agora inbox for [email protected]  

help / color / mirror / Atom feed

[PATCH v3 4/7] Row pattern recognition patch (executor).
201+ messages / 4 participants
[nested] [flat]

* [PATCH v3 4/7] Row pattern recognition patch (executor).
@ 2023-07-26 10:49  Tatsuo Ishii <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Tatsuo Ishii @ 2023-07-26 10:49 UTC (permalink / raw)

---
 src/backend/executor/nodeWindowAgg.c | 701 ++++++++++++++++++++++++++-
 src/backend/utils/adt/windowfuncs.c  |  38 +-
 src/include/catalog/pg_proc.dat      |   6 +
 src/include/nodes/execnodes.h        |  18 +
 src/include/windowapi.h              |   8 +
 5 files changed, 758 insertions(+), 13 deletions(-)

diff --git a/src/backend/executor/nodeWindowAgg.c b/src/backend/executor/nodeWindowAgg.c
index 310ac23e3a..0586bf57d6 100644
--- a/src/backend/executor/nodeWindowAgg.c
+++ b/src/backend/executor/nodeWindowAgg.c
@@ -36,6 +36,7 @@
 #include "access/htup_details.h"
 #include "catalog/objectaccess.h"
 #include "catalog/pg_aggregate.h"
+#include "catalog/pg_collation_d.h"
 #include "catalog/pg_proc.h"
 #include "executor/executor.h"
 #include "executor/nodeWindowAgg.h"
@@ -48,6 +49,7 @@
 #include "utils/acl.h"
 #include "utils/builtins.h"
 #include "utils/datum.h"
+#include "utils/fmgroids.h"
 #include "utils/expandeddatum.h"
 #include "utils/lsyscache.h"
 #include "utils/memutils.h"
@@ -159,6 +161,14 @@ typedef struct WindowStatePerAggData
 	bool		restart;		/* need to restart this agg in this cycle? */
 } WindowStatePerAggData;
 
+/*
+ * Map between Var attno in a target list and the parsed attno.
+ */
+typedef struct AttnoMap {
+	List		*attno;			/* att number in target list (list of AttNumber) */
+	List		*attnosyn;		/* parsed att number (list of AttNumber) */
+} AttnoMap;
+
 static void initialize_windowaggregate(WindowAggState *winstate,
 									   WindowStatePerFunc perfuncstate,
 									   WindowStatePerAgg peraggstate);
@@ -182,8 +192,9 @@ static void begin_partition(WindowAggState *winstate);
 static void spool_tuples(WindowAggState *winstate, int64 pos);
 static void release_partition(WindowAggState *winstate);
 
-static int	row_is_in_frame(WindowAggState *winstate, int64 pos,
+static int  row_is_in_frame(WindowAggState *winstate, int64 pos,
 							TupleTableSlot *slot);
+
 static void update_frameheadpos(WindowAggState *winstate);
 static void update_frametailpos(WindowAggState *winstate);
 static void update_grouptailpos(WindowAggState *winstate);
@@ -195,9 +206,19 @@ static Datum GetAggInitVal(Datum textInitVal, Oid transtype);
 
 static bool are_peers(WindowAggState *winstate, TupleTableSlot *slot1,
 					  TupleTableSlot *slot2);
-static bool window_gettupleslot(WindowObject winobj, int64 pos,
-								TupleTableSlot *slot);
 
+static void attno_map(Node *node, AttnoMap *map);
+static bool attno_map_walker(Node *node, void *context);
+static int row_is_in_reduced_frame(WindowObject winobj, int64 pos);
+
+static int64 evaluate_pattern(WindowObject winobj, int64 current_pos, 
+							  char *vname, StringInfo encoded_str, bool *result);
+
+static bool get_slots(WindowObject winobj, int64 current_pos);
+
+static int search_str_set(char *pattern, StringInfo *str_set, int set_size);
+static void search_str_set_recurse(char *pattern, StringInfo *str_set, int set_size, int set_index,
+								   char *encoded_str, int *resultlen);
 
 /*
  * initialize_windowaggregate
@@ -673,6 +694,9 @@ eval_windowaggregates(WindowAggState *winstate)
 	WindowObject agg_winobj;
 	TupleTableSlot *agg_row_slot;
 	TupleTableSlot *temp_slot;
+	bool		reduced_frame_set;
+	bool		check_reduced_frame;
+	int			num_rows_in_reduced_frame;
 
 	numaggs = winstate->numaggs;
 	if (numaggs == 0)
@@ -790,6 +814,7 @@ eval_windowaggregates(WindowAggState *winstate)
 			(winstate->frameOptions & FRAMEOPTION_EXCLUSION) ||
 			winstate->aggregatedupto <= winstate->frameheadpos)
 		{
+			elog(DEBUG1, "peraggstate->restart  is set");
 			peraggstate->restart = true;
 			numaggs_restart++;
 		}
@@ -861,8 +886,10 @@ eval_windowaggregates(WindowAggState *winstate)
 	 * If we created a mark pointer for aggregates, keep it pushed up to frame
 	 * head, so that tuplestore can discard unnecessary rows.
 	 */
+#ifdef NOT_USED
 	if (agg_winobj->markptr >= 0)
 		WinSetMarkPosition(agg_winobj, winstate->frameheadpos);
+#endif
 
 	/*
 	 * Now restart the aggregates that require it.
@@ -919,6 +946,10 @@ eval_windowaggregates(WindowAggState *winstate)
 		ExecClearTuple(agg_row_slot);
 	}
 
+	reduced_frame_set = false;
+	check_reduced_frame = false;
+	num_rows_in_reduced_frame = 0;
+
 	/*
 	 * Advance until we reach a row not in frame (or end of partition).
 	 *
@@ -930,12 +961,18 @@ eval_windowaggregates(WindowAggState *winstate)
 	{
 		int			ret;
 
+		elog(DEBUG1, "===== loop in frame starts: " INT64_FORMAT, winstate->aggregatedupto);
+
 		/* Fetch next row if we didn't already */
 		if (TupIsNull(agg_row_slot))
 		{
 			if (!window_gettupleslot(agg_winobj, winstate->aggregatedupto,
 									 agg_row_slot))
+			{
+				if (check_reduced_frame)
+					winstate->aggregatedupto--;
 				break;			/* must be end of partition */
+			}
 		}
 
 		/*
@@ -944,10 +981,47 @@ eval_windowaggregates(WindowAggState *winstate)
 		 */
 		ret = row_is_in_frame(winstate, winstate->aggregatedupto, agg_row_slot);
 		if (ret < 0)
+		{
+			if (winstate->patternVariableList != NIL && check_reduced_frame)
+				winstate->aggregatedupto--;
 			break;
+		}
 		if (ret == 0)
 			goto next_tuple;
 
+		if (winstate->patternVariableList != NIL)
+		{
+			if (!reduced_frame_set)
+			{
+				num_rows_in_reduced_frame = row_is_in_reduced_frame(winstate->agg_winobj, winstate->aggregatedupto);
+				reduced_frame_set = true;
+				elog(DEBUG1, "set num_rows_in_reduced_frame: %d pos: " INT64_FORMAT,
+					 num_rows_in_reduced_frame, winstate->aggregatedupto);
+
+				if (num_rows_in_reduced_frame <= 0)
+					break;
+
+				else if (num_rows_in_reduced_frame > 0)
+					check_reduced_frame = true;
+			}
+
+			if (check_reduced_frame)
+			{
+				elog(DEBUG1, "decrease num_rows_in_reduced_frame: %d pos: " INT64_FORMAT,
+					 num_rows_in_reduced_frame, winstate->aggregatedupto);
+				num_rows_in_reduced_frame--;
+				if (num_rows_in_reduced_frame < 0)
+				{
+					/*
+					 * No more rows remain in the reduced frame. Finish
+					 * accumulating row into the aggregates.
+					 */
+					winstate->aggregatedupto--;
+					break;
+				}
+			}
+		}
+
 		/* Set tuple context for evaluation of aggregate arguments */
 		winstate->tmpcontext->ecxt_outertuple = agg_row_slot;
 
@@ -976,6 +1050,8 @@ next_tuple:
 		ExecClearTuple(agg_row_slot);
 	}
 
+	elog(DEBUG1, "===== break loop in frame starts: " INT64_FORMAT, winstate->aggregatedupto);
+
 	/* The frame's end is not supposed to move backwards, ever */
 	Assert(aggregatedupto_nonrestarted <= winstate->aggregatedupto);
 
@@ -2053,6 +2129,8 @@ ExecWindowAgg(PlanState *pstate)
 
 	CHECK_FOR_INTERRUPTS();
 
+	elog(DEBUG1, "ExecWindowAgg called. pos: " INT64_FORMAT , winstate->currentpos);
+
 	if (winstate->status == WINDOWAGG_DONE)
 		return NULL;
 
@@ -2388,6 +2466,12 @@ ExecInitWindowAgg(WindowAgg *node, EState *estate, int eflags)
 	TupleDesc	scanDesc;
 	ListCell   *l;
 
+	TargetEntry	*te;
+	Expr		*expr;
+	Var			*var;
+	int			nargs;
+	AttnoMap	attnomap;
+
 	/* check for unsupported flags */
 	Assert(!(eflags & (EXEC_FLAG_BACKWARD | EXEC_FLAG_MARK)));
 
@@ -2483,6 +2567,16 @@ ExecInitWindowAgg(WindowAgg *node, EState *estate, int eflags)
 	winstate->temp_slot_2 = ExecInitExtraTupleSlot(estate, scanDesc,
 												   &TTSOpsMinimalTuple);
 
+	winstate->prev_slot = ExecInitExtraTupleSlot(estate, scanDesc,
+												 &TTSOpsMinimalTuple);
+
+	winstate->next_slot = ExecInitExtraTupleSlot(estate, scanDesc,
+												 &TTSOpsMinimalTuple);
+
+	winstate->null_slot = ExecInitExtraTupleSlot(estate, scanDesc,
+												 &TTSOpsMinimalTuple);
+	winstate->null_slot = ExecStoreAllNullTuple(winstate->null_slot);
+
 	/*
 	 * create frame head and tail slots only if needed (must create slots in
 	 * exactly the same cases that update_frameheadpos and update_frametailpos
@@ -2667,6 +2761,69 @@ ExecInitWindowAgg(WindowAgg *node, EState *estate, int eflags)
 	winstate->inRangeAsc = node->inRangeAsc;
 	winstate->inRangeNullsFirst = node->inRangeNullsFirst;
 
+	/* Set up SKIP TO type */
+	winstate->rpSkipTo = node->rpSkipTo;
+	/* Set up row pattern recognition PATTERN clause */
+	winstate->patternVariableList = node->patternVariable;
+	winstate->patternRegexpList = node->patternRegexp;
+
+	/* Set up row pattern recognition DEFINE clause */
+
+	/*
+	 * Collect mapping between varattno and varattnosyn in the targetlist.
+	 * XXX: For now we only check RPR's argument. Eventually we have to
+	 * recurse the targetlist to find out all mappings in Var nodes.
+	 */
+	attnomap.attno = NIL;
+	attnomap.attnosyn = NIL;
+
+	foreach (l, node->plan.targetlist)
+	{
+		te = lfirst(l);
+		if (IsA(te->expr, WindowFunc))
+		{
+			WindowFunc	*func = (WindowFunc *)te->expr;
+
+			/* sanity check */
+			nargs = list_length(func->args);
+			if (nargs != 1)
+				continue;
+
+			expr = (Expr *) lfirst(list_head(func->args));
+			if (!IsA(expr, Var))
+				continue;
+
+			var = (Var *)expr;
+			elog(DEBUG1, "resname: %s varattno: %d varattnosyn: %d",
+				 te->resname, var->varattno, var->varattnosyn);
+			attnomap.attno = lappend_int(attnomap.attno, var->varattno);
+			attnomap.attnosyn = lappend_int(attnomap.attnosyn, var->varattnosyn);
+		}
+	}
+
+	winstate->defineVariableList = NIL;
+	winstate->defineClauseList = NIL;
+	if (node->defineClause != NIL)
+	{
+		foreach(l, node->defineClause)
+		{
+			char		*name;
+			ExprState	*exps;
+
+			te = lfirst(l);
+			name = te->resname;
+			expr = te->expr;
+
+			elog(DEBUG1, "defineVariable name: %s", name);
+			winstate->defineVariableList = lappend(winstate->defineVariableList,
+												   makeString(pstrdup(name)));
+			/* tweak expr so that it referes to outer slot */
+			attno_map((Node *)expr, &attnomap);
+			exps = ExecInitExpr(expr, (PlanState *) winstate);
+			winstate->defineClauseList = lappend(winstate->defineClauseList, exps);
+		}
+	}
+
 	winstate->all_first = true;
 	winstate->partition_spooled = false;
 	winstate->more_partitions = false;
@@ -2674,6 +2831,77 @@ ExecInitWindowAgg(WindowAgg *node, EState *estate, int eflags)
 	return winstate;
 }
 
+/*
+ * Rewrite Var node's varattno to the varattno which is used in the target
+ * list using AttnoMap.  We also rewrite varno so that it sees outer tuple
+ * (PREV) or inner tuple (NEXT).
+ */
+static void
+attno_map(Node *node, AttnoMap *map)
+{
+	(void) expression_tree_walker(node, attno_map_walker, (void *) map);
+}
+
+static bool
+attno_map_walker(Node *node, void *context)
+{
+	FuncExpr	*func;
+	int			nargs;
+	Expr		*expr;
+	Var			*var;
+	AttnoMap	*attnomap;
+	ListCell	*lc1, *lc2;
+	
+	if (node == NULL)
+		return false;
+
+	attnomap = (AttnoMap *) context;
+
+	if (IsA(node, FuncExpr))
+	{
+		func = (FuncExpr *)node;
+
+		if (func->funcid == F_PREV || func->funcid == F_NEXT)
+		{
+			/* sanity check */
+			nargs = list_length(func->args);
+			if (list_length(func->args) != 1)
+				elog(ERROR, "PREV/NEXT must have 1 argument but function %d has %d args", func->funcid, nargs);
+
+			expr = (Expr *) lfirst(list_head(func->args));
+			if (!IsA(expr, Var))
+				elog(ERROR, "PREV/NEXT's arg is not Var");	/* XXX: is it possible that arg type is Const? */
+			var = (Var *)expr;
+
+			if (func->funcid == F_PREV)
+				var->varno = OUTER_VAR;
+			else
+				var->varno = INNER_VAR;
+		}
+		return expression_tree_walker(node, attno_map_walker, (void *) context);
+	}
+	else if (IsA(node, Var))
+	{
+		var = (Var *)node;	 
+
+		elog(DEBUG1, "original varno: %d varattno: %d", var->varno, var->varattno);
+
+		forboth(lc1, attnomap->attno, lc2, attnomap->attnosyn)
+		{
+			int	attno = lfirst_int(lc1);
+			int	attnosyn = lfirst_int(lc2);
+
+			elog(DEBUG1, "walker: varattno: %d varattnosyn: %d",attno, attnosyn);
+			if (var->varattno == attnosyn)
+			{
+				elog(DEBUG1, "loc: %d rewrite varattno from: %d to %d", var->location, attnosyn, attno);
+				var->varattno = attno;
+			}
+		}
+	}
+	return expression_tree_walker(node, attno_map_walker, (void *) context);
+}
+
 /* -----------------
  * ExecEndWindowAgg
  * -----------------
@@ -2691,6 +2919,8 @@ ExecEndWindowAgg(WindowAggState *node)
 	ExecClearTuple(node->agg_row_slot);
 	ExecClearTuple(node->temp_slot_1);
 	ExecClearTuple(node->temp_slot_2);
+	ExecClearTuple(node->prev_slot);
+	ExecClearTuple(node->next_slot);
 	if (node->framehead_slot)
 		ExecClearTuple(node->framehead_slot);
 	if (node->frametail_slot)
@@ -2740,6 +2970,8 @@ ExecReScanWindowAgg(WindowAggState *node)
 	ExecClearTuple(node->agg_row_slot);
 	ExecClearTuple(node->temp_slot_1);
 	ExecClearTuple(node->temp_slot_2);
+	ExecClearTuple(node->prev_slot);
+	ExecClearTuple(node->next_slot);
 	if (node->framehead_slot)
 		ExecClearTuple(node->framehead_slot);
 	if (node->frametail_slot)
@@ -3080,7 +3312,7 @@ are_peers(WindowAggState *winstate, TupleTableSlot *slot1,
  *
  * Returns true if successful, false if no such row
  */
-static bool
+bool
 window_gettupleslot(WindowObject winobj, int64 pos, TupleTableSlot *slot)
 {
 	WindowAggState *winstate = winobj->winstate;
@@ -3100,7 +3332,7 @@ window_gettupleslot(WindowObject winobj, int64 pos, TupleTableSlot *slot)
 		return false;
 
 	if (pos < winobj->markpos)
-		elog(ERROR, "cannot fetch row before WindowObject's mark position");
+		elog(ERROR, "cannot fetch row: " INT64_FORMAT " before WindowObject's mark position: " INT64_FORMAT,  pos, winobj->markpos );
 
 	oldcontext = MemoryContextSwitchTo(winstate->ss.ps.ps_ExprContext->ecxt_per_query_memory);
 
@@ -3420,14 +3652,54 @@ WinGetFuncArgInFrame(WindowObject winobj, int argno,
 	WindowAggState *winstate;
 	ExprContext *econtext;
 	TupleTableSlot *slot;
-	int64		abs_pos;
-	int64		mark_pos;
 
 	Assert(WindowObjectIsValid(winobj));
 	winstate = winobj->winstate;
 	econtext = winstate->ss.ps.ps_ExprContext;
 	slot = winstate->temp_slot_1;
 
+	if (WinGetSlotInFrame(winobj, slot,
+						  relpos, seektype, set_mark,
+						  isnull, isout) == 0)
+	{
+		econtext->ecxt_outertuple = slot;
+		return ExecEvalExpr((ExprState *) list_nth(winobj->argstates, argno),
+							econtext, isnull);
+	}
+
+	if (isout)
+		*isout = true;
+	*isnull = true;
+	return (Datum) 0;
+}
+
+/*
+ * WinGetSlotInFrame
+ * slot: TupleTableSlot to store the result
+ * relpos: signed rowcount offset from the seek position
+ * seektype: WINDOW_SEEK_HEAD or WINDOW_SEEK_TAIL
+ * set_mark: If the row is found/in frame and set_mark is true, the mark is
+ *		moved to the row as a side-effect.
+ * isnull: output argument, receives isnull status of result
+ * isout: output argument, set to indicate whether target row position
+ *		is out of frame (can pass NULL if caller doesn't care about this)
+ *
+ * Returns 0 if we successfullt got the slot. false if out of frame.
+ * (also isout is set)
+ */
+int
+WinGetSlotInFrame(WindowObject winobj, TupleTableSlot *slot,
+					 int relpos, int seektype, bool set_mark,
+					 bool *isnull, bool *isout)
+{
+	WindowAggState *winstate;
+	int64		abs_pos;
+	int64		mark_pos;
+	int			num_reduced_frame;
+
+	Assert(WindowObjectIsValid(winobj));
+	winstate = winobj->winstate;
+
 	switch (seektype)
 	{
 		case WINDOW_SEEK_CURRENT:
@@ -3494,6 +3766,12 @@ WinGetFuncArgInFrame(WindowObject winobj, int argno,
 						 winstate->frameOptions);
 					break;
 			}
+			num_reduced_frame = row_is_in_reduced_frame(winobj, winstate->frameheadpos);
+			if (num_reduced_frame < 0)
+				goto out_of_frame;
+			else if (num_reduced_frame > 0)
+				if (relpos >= num_reduced_frame)
+					goto out_of_frame;
 			break;
 		case WINDOW_SEEK_TAIL:
 			/* rejecting relpos > 0 is easy and simplifies code below */
@@ -3565,6 +3843,12 @@ WinGetFuncArgInFrame(WindowObject winobj, int argno,
 					mark_pos = 0;	/* keep compiler quiet */
 					break;
 			}
+
+			num_reduced_frame = row_is_in_reduced_frame(winobj, winstate->frameheadpos + relpos);
+			if (num_reduced_frame < 0)
+				goto out_of_frame;
+			else if (num_reduced_frame > 0)
+				abs_pos = winstate->frameheadpos + relpos + num_reduced_frame - 1;
 			break;
 		default:
 			elog(ERROR, "unrecognized window seek type: %d", seektype);
@@ -3583,15 +3867,13 @@ WinGetFuncArgInFrame(WindowObject winobj, int argno,
 		*isout = false;
 	if (set_mark)
 		WinSetMarkPosition(winobj, mark_pos);
-	econtext->ecxt_outertuple = slot;
-	return ExecEvalExpr((ExprState *) list_nth(winobj->argstates, argno),
-						econtext, isnull);
+	return 0;
 
 out_of_frame:
 	if (isout)
 		*isout = true;
 	*isnull = true;
-	return (Datum) 0;
+	return -1;
 }
 
 /*
@@ -3622,3 +3904,400 @@ WinGetFuncArgCurrent(WindowObject winobj, int argno, bool *isnull)
 	return ExecEvalExpr((ExprState *) list_nth(winobj->argstates, argno),
 						econtext, isnull);
 }
+
+WindowAggState *
+WinGetAggState(WindowObject winobj)
+{
+	return winobj->winstate;
+}
+
+/*
+ * row_is_in_reduced_frame
+ * Determine whether a row is in the current row's reduced window frame according
+ * to row pattern matching
+ *
+ * The row must has been already determined that it is in a full window frame
+ * and fetched it into slot.
+ *
+ * Returns:
+ * = 0, RPR is not defined.
+ * >0, if the row is the first in the reduced frame. Return the number of rows in the reduced frame.
+ * -1, if the row is unmatched row
+ * -2, if the row is in the reduced frame but needed to be skipped because of
+ * AFTER MATCH SKIP PAST LAST ROW
+ */
+static
+int row_is_in_reduced_frame(WindowObject winobj, int64 pos)
+{
+	WindowAggState *winstate = winobj->winstate;
+	ListCell	*lc1, *lc2;
+	bool		expression_result;
+	int			num_matched_rows;
+	int64		original_pos;
+	bool		anymatch;
+	StringInfo	encoded_str;
+	StringInfo	pattern_str = makeStringInfo();
+
+	/*
+	 * Array of pattern variables evaluted to true.
+	 * Each character corresponds to pattern variable.
+	 * Example:
+	 * str_set[0] = "AB";
+	 * str_set[1] = "AC";
+	 * In this case at row 0 A and B are true, and A and C are true in row 1.
+	 */
+	#define ENCODED_STR_ARRAY_ALLOC_SIZE	128
+	StringInfo	*str_set = NULL;
+	int		str_set_index;
+	int		str_set_size;
+
+	if (winstate->patternVariableList == NIL)
+	{
+		/*
+		 * RPR is not defined. Assume that we are always in the the reduced
+		 * window frame.
+		 */
+		return 0;
+	}
+
+	/* save original pos */
+	original_pos = pos;
+
+	/*
+	 * Check whether the row speicied by pos is in the reduced frame. The
+	 * second and subsequent rows need to be recognized as "unmatched" rows if
+	 * AFTER MATCH SKIP PAST LAST ROW is defined.
+	 */
+	if (winstate->rpSkipTo == ST_PAST_LAST_ROW &&
+		pos > winstate->headpos_in_reduced_frame &&
+		pos < (winstate->headpos_in_reduced_frame + winstate->num_rows_in_reduced_frame))
+		return -2;
+		
+	/*
+	 * Loop over until none of pattern matches or encounters end of frame.
+	 */
+	for (;;)
+	{
+		int64	result_pos = -1;
+
+		/*
+		 * Loop over each PATTERN variable.
+		 */
+		anymatch = false;
+		encoded_str = makeStringInfo();
+
+		forboth(lc1, winstate->patternVariableList, lc2, winstate->patternRegexpList)
+		{
+			char	*vname = strVal(lfirst(lc1));
+			char	*quantifier = strVal(lfirst(lc2));
+
+			elog(DEBUG1, "pos: " INT64_FORMAT " pattern vname: %s quantifier: %s", pos, vname, quantifier);
+
+			expression_result = false;
+
+			/* evaluate row pattern against current row */
+			result_pos = evaluate_pattern(winobj, pos, vname, encoded_str, &expression_result);
+			if (expression_result)
+			{
+				elog(DEBUG1, "expression result is true");
+				anymatch = true;
+			}
+
+			/*
+			 * If out of frame, we are done.
+			 */
+			 if (result_pos < 0)
+				 break;
+		}
+
+		if (!anymatch)
+		{
+			/* none of patterns matched. */
+			break;
+		}
+
+		/* build encoded string array */
+		if (str_set == NULL)
+		{
+			str_set_index = 0;
+			str_set_size = ENCODED_STR_ARRAY_ALLOC_SIZE * sizeof(StringInfo);
+			str_set = palloc(str_set_size);
+		}
+
+		str_set[str_set_index++] = encoded_str;
+
+		elog(DEBUG1, "pos: " INT64_FORMAT " str_set_index: %d encoded_str: %s", pos, str_set_index, encoded_str->data);
+
+		if (str_set_index >= str_set_size)
+		{
+			str_set_size *= 2;
+			str_set = repalloc(str_set, str_set_size);
+		}
+
+		/* move to next row */
+		pos++;
+
+		if (result_pos < 0)
+		{
+			/* out of frame */
+			break;
+		}
+	}
+
+	if (str_set == NULL)
+	{
+		/* no matches found in the first row */
+		return -1;
+	}
+
+	elog(DEBUG1, "pos: " INT64_FORMAT " encoded_str: %s", pos, encoded_str->data);
+
+	/* build regular expression */
+	pattern_str = makeStringInfo();
+	appendStringInfoChar(pattern_str, '^');
+	forboth(lc1, winstate->patternVariableList, lc2, winstate->patternRegexpList)
+	{
+		char	*vname = strVal(lfirst(lc1));
+		char	*quantifier = strVal(lfirst(lc2));
+
+		appendStringInfoChar(pattern_str, vname[0]);
+		if (quantifier[0])
+			appendStringInfoChar(pattern_str, quantifier[0]);
+		elog(DEBUG1, "vname: %s quantifier: %s", vname, quantifier);
+	}
+
+	elog(DEBUG1, "pos: " INT64_FORMAT " pattern: %s", pos, pattern_str->data);
+
+	/* look for matching pattern variable sequence */
+	num_matched_rows = search_str_set(pattern_str->data, str_set, str_set_index);
+	if (num_matched_rows <= 0)
+		return -1;
+
+	/*
+	 * We are at the first row in the reduced frame.  Save the number of
+	 * matched rows as the number of rows in the reduced frame.
+	 */
+	winstate->headpos_in_reduced_frame = original_pos;
+	winstate->num_rows_in_reduced_frame = num_matched_rows;
+
+	return num_matched_rows;
+}
+
+/*
+ * search set of encode_str.
+ * set_size: size of set_str array.
+ */
+static
+int search_str_set(char *pattern, StringInfo *str_set, int set_size)
+{
+	char		*encoded_str = palloc0(set_size+1);
+	int			resultlen = 0;
+
+	search_str_set_recurse(pattern, str_set, set_size, 0, encoded_str, &resultlen);
+	elog(DEBUG1, "search_str_set returns %d", resultlen);
+	return resultlen;
+}
+
+static
+void search_str_set_recurse(char *pattern, StringInfo *str_set, int set_size, int set_index, char *encoded_str, int *resultlen)
+{
+	char	*p;
+
+	if (set_index >= set_size)
+	{
+		Datum	d;
+		text	*res;
+		char	*substr;
+
+		/*
+		 * We first perform pattern matching using regexp_instr, then call
+		 * textregexsubstr to get matched substring to know how log the
+		 * matched string is. That is the number of rows in the reduced window
+		 * frame.  The reason why we can't call textregexsubstr is, it error
+		 * out if pattern is not match.
+		 */
+		if (DatumGetInt32(DirectFunctionCall2Coll(regexp_instr, DEFAULT_COLLATION_OID,
+												  PointerGetDatum(cstring_to_text(encoded_str)),
+												  PointerGetDatum(cstring_to_text(pattern)))) > 0)
+		{
+			d = DirectFunctionCall2Coll(textregexsubstr,
+										DEFAULT_COLLATION_OID,
+										PointerGetDatum(cstring_to_text(encoded_str)),
+										PointerGetDatum(cstring_to_text(pattern)));
+			if (d != 0)
+			{
+				int		len;
+
+				res = DatumGetTextPP(d);
+				substr = text_to_cstring(res);
+				len = strlen(substr);
+				if (len > *resultlen)
+					/* remember the longest match */
+					*resultlen = len;
+			}
+		}
+		return;
+	}
+
+	p = str_set[set_index]->data;
+	while (*p)
+	{
+		encoded_str[set_index] = *p;
+		p++;
+		search_str_set_recurse(pattern, str_set, set_size, set_index + 1, encoded_str, resultlen);
+	}
+}
+
+
+/*
+ * Evaluate expression associated with PATTERN variable vname.
+ * relpos is relative row position in a frame (starting from 0).
+ * "quantifier" is the quatifier part of the PATTERN regular expression.
+ * Currently only '+' is allowed.
+ * result is out paramater representing the expression evaluation result
+ * is true of false.
+ * Return values are:
+ * >=0: the last match absolute row position
+ * other wise out of frame.
+ */
+static
+int64 evaluate_pattern(WindowObject winobj, int64 current_pos, 
+						char *vname, StringInfo encoded_str, bool *result)
+{
+	WindowAggState	*winstate = winobj->winstate;
+	ExprContext		*econtext = winstate->ss.ps.ps_ExprContext;
+	ListCell		*lc1, *lc2;
+	ExprState		*pat;
+	Datum			eval_result;
+	bool			out_of_frame = false;
+	bool			isnull;
+
+	forboth (lc1, winstate->defineVariableList, lc2, winstate->defineClauseList)
+	{
+		char	*name = strVal(lfirst(lc1));
+
+		if (strcmp(vname, name))
+			continue;
+
+		/* set expression to evaluate */
+		pat = lfirst(lc2);
+
+		/* get current, previous and next tuples */
+		if (!get_slots(winobj, current_pos))
+		{
+			out_of_frame = true;
+		}
+		else
+		{
+			/* evaluate the expression */
+			eval_result = ExecEvalExpr(pat, econtext, &isnull);
+			if (isnull)
+			{
+				/* expression is NULL */
+				elog(DEBUG1, "expression for %s is NULL at row: " INT64_FORMAT, vname, current_pos);
+				*result = false;
+			}
+			else
+			{
+				if (!DatumGetBool(eval_result))
+				{
+					/* expression is false */
+					elog(DEBUG1, "expression for %s is false at row: " INT64_FORMAT, vname, current_pos);
+					*result = false;
+				}
+				else
+				{
+					/* expression is true */
+					elog(DEBUG1, "expression for %s is true at row: " INT64_FORMAT, vname, current_pos);
+					appendStringInfoChar(encoded_str, vname[0]);
+					*result = true;
+				}
+			}
+			break;
+		}
+
+		if (out_of_frame)
+		{
+			*result = false;
+			return -1;
+		}
+	}
+	return current_pos;
+}
+
+/*
+ * Get current, previous and next tuples.
+ * Returns false if current row is out of partition/full frame.
+ */
+static
+bool get_slots(WindowObject winobj, int64 current_pos)
+{
+	WindowAggState *winstate = winobj->winstate;
+	TupleTableSlot *slot;
+	int		ret;
+	ExprContext *econtext;
+
+	econtext = winstate->ss.ps.ps_ExprContext;
+
+	/* set up current row tuple slot */
+	slot = winstate->temp_slot_1;
+	if (!window_gettupleslot(winobj, current_pos, slot))
+	{
+		elog(DEBUG1, "current row is out of partition at:" INT64_FORMAT, current_pos);
+		return false;
+
+		ret = row_is_in_frame(winstate, current_pos, slot);
+		if (ret <= 0)
+		{
+			elog(DEBUG1, "current row is out of frame at: " INT64_FORMAT, current_pos);
+			return false;
+		}
+	}
+	econtext->ecxt_scantuple = slot;
+
+	/* for PREV */
+	if (current_pos > 0)
+	{
+		slot = winstate->prev_slot;
+		if (!window_gettupleslot(winobj, current_pos - 1, slot))
+		{
+			elog(DEBUG1, "previous row is out of partition at: " INT64_FORMAT, current_pos - 1);
+			econtext->ecxt_outertuple = winstate->null_slot;
+		}
+		else
+		{
+			ret = row_is_in_frame(winstate, current_pos - 1, slot);
+			if (ret <= 0)
+			{
+				elog(DEBUG1, "previous row is out of frame at: " INT64_FORMAT, current_pos - 1);
+				econtext->ecxt_outertuple = winstate->null_slot;
+			}
+			else
+			{
+				econtext->ecxt_outertuple = slot;
+			}
+		}
+	}
+	else
+		econtext->ecxt_outertuple = winstate->null_slot;
+
+	/* for NEXT */
+	slot = winstate->next_slot;
+	if (!window_gettupleslot(winobj, current_pos + 1, slot))
+	{
+		elog(DEBUG1, "next row is out of partiton at: " INT64_FORMAT, current_pos + 1);
+		econtext->ecxt_innertuple = winstate->null_slot;
+	}
+	else
+	{
+		ret = row_is_in_frame(winstate, current_pos + 1, slot);
+		if (ret <= 0)
+		{
+			elog(DEBUG1, "next row is out of frame at: " INT64_FORMAT, current_pos + 1);
+			econtext->ecxt_innertuple = winstate->null_slot;
+		}
+		else
+			econtext->ecxt_innertuple = slot;
+	}
+	return true;
+}
diff --git a/src/backend/utils/adt/windowfuncs.c b/src/backend/utils/adt/windowfuncs.c
index b87a624fb2..e4cab36ec9 100644
--- a/src/backend/utils/adt/windowfuncs.c
+++ b/src/backend/utils/adt/windowfuncs.c
@@ -13,6 +13,9 @@
  */
 #include "postgres.h"
 
+#include "catalog/pg_collation_d.h"
+#include "executor/executor.h"
+#include "nodes/execnodes.h"
 #include "nodes/supportnodes.h"
 #include "utils/builtins.h"
 #include "windowapi.h"
@@ -36,11 +39,19 @@ typedef struct
 	int64		remainder;		/* (total rows) % (bucket num) */
 } ntile_context;
 
+/*
+ * rpr process information.
+ * Used for AFTER MATCH SKIP PAST LAST ROW
+ */
+typedef struct SkipContext
+{
+	int64		pos;	/* last row absolute position */
+} SkipContext;
+
 static bool rank_up(WindowObject winobj);
 static Datum leadlag_common(FunctionCallInfo fcinfo,
 							bool forward, bool withoffset, bool withdefault);
 
-
 /*
  * utility routine for *_rank functions.
  */
@@ -673,7 +684,7 @@ window_last_value(PG_FUNCTION_ARGS)
 	bool		isnull;
 
 	result = WinGetFuncArgInFrame(winobj, 0,
-								  0, WINDOW_SEEK_TAIL, true,
+								  0, WINDOW_SEEK_TAIL, false,
 								  &isnull, NULL);
 	if (isnull)
 		PG_RETURN_NULL();
@@ -713,3 +724,26 @@ window_nth_value(PG_FUNCTION_ARGS)
 
 	PG_RETURN_DATUM(result);
 }
+
+/*
+ * prev
+ * Dummy function to invoke RPR's navigation operator "PREV".
+ * This is *not* a window function.
+ */
+Datum
+window_prev(PG_FUNCTION_ARGS)
+{
+	PG_RETURN_DATUM(PG_GETARG_DATUM(0));
+}
+
+/*
+ * next
+ * Dummy function to invoke RPR's navigation operation "NEXT".
+ * This is *not* a window function.
+ */
+Datum
+window_next(PG_FUNCTION_ARGS)
+{
+	PG_RETURN_DATUM(PG_GETARG_DATUM(0));
+}
+
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 6996073989..fa100b2665 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10397,6 +10397,12 @@
 { oid => '3114', descr => 'fetch the Nth row value',
   proname => 'nth_value', prokind => 'w', prorettype => 'anyelement',
   proargtypes => 'anyelement int4', prosrc => 'window_nth_value' },
+{ oid => '6122', descr => 'previous value',
+  proname => 'prev', provolatile => 's', prorettype => 'anyelement',
+  proargtypes => 'anyelement', prosrc => 'window_prev' },
+{ oid => '6123', descr => 'next value',
+  proname => 'next', provolatile => 's', prorettype => 'anyelement',
+  proargtypes => 'anyelement', prosrc => 'window_next' },
 
 # functions for range types
 { oid => '3832', descr => 'I/O',
diff --git a/src/include/nodes/execnodes.h b/src/include/nodes/execnodes.h
index cb714f4a19..4fd3bd1a93 100644
--- a/src/include/nodes/execnodes.h
+++ b/src/include/nodes/execnodes.h
@@ -2519,6 +2519,14 @@ typedef struct WindowAggState
 	int64		groupheadpos;	/* current row's peer group head position */
 	int64		grouptailpos;	/* " " " " tail position (group end+1) */
 
+	/* these fields are used in Row pattern recognition: */
+	RPSkipTo	rpSkipTo;		/* Row Pattern Skip To type */	
+	List	   *patternVariableList;	/* list of row pattern variables names (list of String) */
+	List	   *patternRegexpList;	/* list of row pattern regular expressions ('+' or ''. list of String) */
+	List	   *defineVariableList;	/* list of row pattern definition variables (list of String) */
+	List	   *defineClauseList;	/* expression for row pattern definition
+									 * search conditions ExprState list */
+
 	MemoryContext partcontext;	/* context for partition-lifespan data */
 	MemoryContext aggcontext;	/* shared context for aggregate working data */
 	MemoryContext curaggcontext;	/* current aggregate's working data */
@@ -2555,6 +2563,16 @@ typedef struct WindowAggState
 	TupleTableSlot *agg_row_slot;
 	TupleTableSlot *temp_slot_1;
 	TupleTableSlot *temp_slot_2;
+
+	/* temporary slots for RPR */
+	TupleTableSlot *prev_slot;	/* PREV row navigation operator */
+	TupleTableSlot *next_slot;	/* NEXT row navigation operator */
+	TupleTableSlot *null_slot;	/* all NULL slot */
+
+	/* head of the reduced window frame */
+	int64		headpos_in_reduced_frame;
+	/* number of rows in the reduced window frame */
+	int64		num_rows_in_reduced_frame;
 } WindowAggState;
 
 /* ----------------
diff --git a/src/include/windowapi.h b/src/include/windowapi.h
index b8c2c565d1..1e292648e9 100644
--- a/src/include/windowapi.h
+++ b/src/include/windowapi.h
@@ -58,7 +58,15 @@ extern Datum WinGetFuncArgInFrame(WindowObject winobj, int argno,
 								  int relpos, int seektype, bool set_mark,
 								  bool *isnull, bool *isout);
 
+extern int WinGetSlotInFrame(WindowObject winobj, TupleTableSlot *slot,
+							 int relpos, int seektype, bool set_mark,
+							 bool *isnull, bool *isout);
+
 extern Datum WinGetFuncArgCurrent(WindowObject winobj, int argno,
 								  bool *isnull);
 
+extern WindowAggState *WinGetAggState(WindowObject winobj);
+
+extern bool window_gettupleslot(WindowObject winobj, int64 pos, TupleTableSlot *slot);
+
 #endif							/* WINDOWAPI_H */
-- 
2.25.1


----Next_Part(Wed_Jul_26_21_21_34_2023_317)--
Content-Type: Text/X-Patch; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="v3-0005-Row-pattern-recognition-patch-docs.patch"



^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread

* [PATCH v1] Add per-backend AIO statistics
@ 2026-06-11 09:45  Bertrand Drouvot <[email protected]>
  0 siblings, 0 replies; 201+ messages in thread

From: Bertrand Drouvot @ 2026-06-11 09:45 UTC (permalink / raw)

This commit adds per-backend AIO statistics, providing per-backend AIO behavior
details.

This data can be retrieved with a new system function called
pg_stat_get_backend_aio(), that returns the following counters based on the PID
provided in input:

- started: total number of AIO operations initiated
- executed_sync: IOs that were executed synchronously (fallback path)
- executed_async: IOs that were submitted asynchronously to the IO method
- completed_self: IO completions processed by the issuing backend itself
- completed_other: IO completions processed on behalf of another backend
- handle_waits: times the backend had to wait for a free AIO handle
- submitted: number of submitted calls to the IO method

These counters are useful for understanding and tuning AIO behavior:

- executed_async / started. A ratio near zero means the backend is falling back
to synchronous execution (TOAST chunk fetches, temp buffers, ...).

- a non-zero handle_waits means the backend exhausted all its AIO handles. That
could mean that io_max_concurrency is too low.

- completed_self vs completed_other reveals cross-backend completion
patterns. That helps see how IO completion work is distributed and could
help interpret per backend IO statistics values.

- executed_async / submitted gives the average batch size per submit call.

This commit is straight-forward, relying on the infrastructure provided
by 9aea73fc61d4 (backend-level pgstats).

XXX: Bump catalog version. No need to touch PGSTAT_FILE_FORMAT_ID as backend
statistics are never written to disk.

Author: Bertrand Drouvot <[email protected]>
Reviewed-by:
Discussion:
---
 doc/src/sgml/monitoring.sgml                |  70 +++++++++++++
 src/backend/storage/aio/aio.c               |  13 +++
 src/backend/utils/activity/pgstat_backend.c | 109 ++++++++++++++++++++
 src/backend/utils/adt/pgstatfuncs.c         |  60 +++++++++++
 src/include/catalog/pg_proc.dat             |   7 ++
 src/include/pgstat.h                        |  25 +++++
 src/include/utils/pgstat_internal.h         |   3 +-
 src/test/modules/test_aio/t/001_aio.pl      |  32 ++++++
 src/tools/pgindent/typedefs.list            |   1 +
 9 files changed, 319 insertions(+), 1 deletion(-)
  24.4% doc/src/sgml/
   3.2% src/backend/storage/aio/
  25.3% src/backend/utils/activity/
  20.5% src/backend/utils/adt/
   4.9% src/include/catalog/
  10.0% src/include/
  11.1% src/test/modules/test_aio/t/

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 12b9ee20d4a..1c566cf0e0e 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -5641,6 +5641,76 @@ description | Waiting for a newly initialized WAL file to reach durable storage
        </para></entry>
       </row>
 
+      <row>
+       <entry id="pg-stat-get-backend-aio" role="func_table_entry"><para role="func_signature">
+        <indexterm>
+         <primary>pg_stat_get_backend_aio</primary>
+        </indexterm>
+        <function>pg_stat_get_backend_aio</function> ( <type>integer</type> )
+        <returnvalue>record</returnvalue>
+       </para>
+       <para>
+        Returns <acronym>AIO</acronym> (Asynchronous I/O) statistics about the
+        backend with the specified process ID. The returned values are:
+       <itemizedlist>
+        <listitem>
+         <para>
+          <literal>started</literal>: Total IOs initiated.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+           <literal>executed_sync</literal>: IOs executed synchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>executed_async</literal>: IOs submitted asynchronously.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_self</literal>: IO completions processed by this
+          backend.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>completed_other</literal>: IO completions processed on
+          behalf of other backends.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>handle_waits</literal>: Times waited for a free AIO handle.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>submitted</literal>: Number of submit calls to the IO
+          method. Compare with <literal>executed_async</literal> to determine
+          average batch size.
+         </para>
+        </listitem>
+        <listitem>
+         <para>
+          <literal>stats_reset</literal>: Timestamp of last stats reset.
+         </para>
+        </listitem>
+       </itemizedlist>
+       </para>
+       <para>
+        The <literal>completed_other</literal> column is only meaningful
+        when <varname>io_method</varname> is set to <literal>io_uring</literal>;
+        with <literal>worker</literal> mode, IO completions are processed by
+        IO worker processes which do not track these statistics.
+       </para>
+       <para>
+        The function does not return AIO statistics for the checkpointer,
+        the background writer, the startup process and the autovacuum launcher.
+       </para></entry>
+      </row>
+
       <row>
        <entry role="func_table_entry"><para role="func_signature">
         <indexterm>
diff --git a/src/backend/storage/aio/aio.c b/src/backend/storage/aio/aio.c
index 8f7e26607b9..507f1727d41 100644
--- a/src/backend/storage/aio/aio.c
+++ b/src/backend/storage/aio/aio.c
@@ -40,6 +40,7 @@
 
 #include "lib/ilist.h"
 #include "miscadmin.h"
+#include "pgstat.h"
 #include "port/atomics.h"
 #include "storage/aio.h"
 #include "storage/aio_internal.h"
@@ -458,6 +459,8 @@ pgaio_io_stage(PgAioHandle *ioh, PgAioOp op)
 				   "staged (synchronous: %d, in_batch: %d)",
 				   needs_synchronous, pgaio_my_backend->in_batchmode);
 
+	pgstat_count_backend_aio_start(needs_synchronous);
+
 	if (!needs_synchronous)
 	{
 		pgaio_my_backend->staged_ios[pgaio_my_backend->num_staged_ios++] = ioh;
@@ -544,6 +547,12 @@ pgaio_io_process_completion(PgAioHandle *ioh, int result)
 	/* condition variable broadcast ensures state is visible before wakeup */
 	ConditionVariableBroadcast(&ioh->cv);
 
+	/* Track AIO completion stats */
+	if (ioh->owner_procno == MyProcNumber)
+		pgstat_count_backend_aio_complete_self();
+	else
+		pgstat_count_backend_aio_complete_other();
+
 	/* contains call to pgaio_io_call_complete_local() */
 	if (ioh->owner_procno == MyProcNumber)
 		pgaio_io_reclaim(ioh);
@@ -762,6 +771,8 @@ pgaio_io_wait_for_free(void)
 {
 	int			reclaimed = 0;
 
+	pgstat_count_backend_aio_handle_wait();
+
 	pgaio_debug(DEBUG2, "waiting for free IO with %d pending, %u in-flight, %u idle IOs",
 				pgaio_my_backend->num_staged_ios,
 				dclist_count(&pgaio_my_backend->in_flight_ios),
@@ -1150,6 +1161,8 @@ pgaio_submit_staged(void)
 
 	Assert(total_submitted == did_submit);
 
+	pgstat_count_backend_aio_submitted();
+
 	pgaio_my_backend->num_staged_ios = 0;
 
 	pgaio_debug(DEBUG4,
diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c
index b736b2ccc6f..4ad755f7f18 100644
--- a/src/backend/utils/activity/pgstat_backend.c
+++ b/src/backend/utils/activity/pgstat_backend.c
@@ -40,6 +40,7 @@
 static PgStat_BackendPending PendingBackendStats;
 static bool backend_has_iostats = false;
 static bool backend_has_lockstats = false;
+static bool backend_has_aiostats = false;
 
 /*
  * WAL usage counters saved from pgWalUsage at the previous call to
@@ -120,6 +121,74 @@ pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type)
 	pgstat_report_fixed = true;
 }
 
+/*
+ * Utility routines to report AIO stats for backends, kept here to avoid
+ * exposing PendingBackendStats to the outside world.
+ */
+void
+pgstat_count_backend_aio_start(bool synchronous)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.started++;
+	if (synchronous)
+		PendingBackendStats.aio_counters.executed_sync++;
+	else
+		PendingBackendStats.aio_counters.executed_async++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_self(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_self++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_complete_other(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.completed_other++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_handle_wait(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.handle_waits++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
+void
+pgstat_count_backend_aio_submitted(void)
+{
+	if (!pgstat_tracks_backend_bktype(MyBackendType))
+		return;
+
+	PendingBackendStats.aio_counters.submitted++;
+
+	backend_has_aiostats = true;
+	pgstat_report_fixed = true;
+}
+
 /*
  * Returns statistics of a backend by proc number.
  */
@@ -326,6 +395,38 @@ pgstat_flush_backend_entry_lock(PgStat_EntryRef *entry_ref)
 	backend_has_lockstats = false;
 }
 
+/*
+ * Flush out locally pending backend AIO statistics.  Locking is managed
+ * by the caller.
+ */
+static void
+pgstat_flush_backend_entry_aio(PgStat_EntryRef *entry_ref)
+{
+	PgStatShared_Backend *shbackendent;
+	PgStat_AioCounters *bktype_shstats;
+
+	if (!backend_has_aiostats)
+		return;
+
+	shbackendent = (PgStatShared_Backend *) entry_ref->shared_stats;
+	bktype_shstats = &shbackendent->stats.aio_counters;
+
+#define AIOSTAT_ACC(fld) \
+	(bktype_shstats->fld += PendingBackendStats.aio_counters.fld)
+	AIOSTAT_ACC(started);
+	AIOSTAT_ACC(executed_sync);
+	AIOSTAT_ACC(executed_async);
+	AIOSTAT_ACC(completed_self);
+	AIOSTAT_ACC(completed_other);
+	AIOSTAT_ACC(handle_waits);
+	AIOSTAT_ACC(submitted);
+#undef AIOSTAT_ACC
+
+	MemSet(&PendingBackendStats.aio_counters, 0, sizeof(PgStat_AioCounters));
+
+	backend_has_aiostats = false;
+}
+
 /*
  * Flush out locally pending backend statistics
  *
@@ -354,6 +455,10 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if ((flags & PGSTAT_BACKEND_FLUSH_LOCK) && backend_has_lockstats)
 		has_pending_data = true;
 
+	/* Some AIO data pending? */
+	if ((flags & PGSTAT_BACKEND_FLUSH_AIO) && backend_has_aiostats)
+		has_pending_data = true;
+
 	if (!has_pending_data)
 		return false;
 
@@ -372,6 +477,9 @@ pgstat_flush_backend(bool nowait, uint32 flags)
 	if (flags & PGSTAT_BACKEND_FLUSH_LOCK)
 		pgstat_flush_backend_entry_lock(entry_ref);
 
+	if (flags & PGSTAT_BACKEND_FLUSH_AIO)
+		pgstat_flush_backend_entry_aio(entry_ref);
+
 	pgstat_unlock_entry(entry_ref);
 
 	return false;
@@ -411,6 +519,7 @@ pgstat_create_backend(ProcNumber procnum)
 	MemSet(&PendingBackendStats, 0, sizeof(PgStat_BackendPending));
 	backend_has_iostats = false;
 	backend_has_lockstats = false;
+	backend_has_aiostats = false;
 
 	/*
 	 * Initialize prevBackendWalUsage with pgWalUsage so that
diff --git a/src/backend/utils/adt/pgstatfuncs.c b/src/backend/utils/adt/pgstatfuncs.c
index 565d0e70768..fb62a56f3fe 100644
--- a/src/backend/utils/adt/pgstatfuncs.c
+++ b/src/backend/utils/adt/pgstatfuncs.c
@@ -1722,6 +1722,66 @@ pg_stat_get_backend_wal(PG_FUNCTION_ARGS)
 	return (pg_stat_wal_build_tuple(bktype_stats, backend_stats->stat_reset_timestamp));
 }
 
+/*
+ * Returns AIO statistics for a backend with given PID.
+ */
+Datum
+pg_stat_get_backend_aio(PG_FUNCTION_ARGS)
+{
+#define PG_STAT_BACKEND_AIO_COLS	8
+	TupleDesc	tupdesc;
+	Datum		values[PG_STAT_BACKEND_AIO_COLS] = {0};
+	bool		nulls[PG_STAT_BACKEND_AIO_COLS] = {0};
+	int			pid;
+	PgStat_Backend *backend_stats;
+	PgStat_AioCounters aio_counters;
+
+	pid = PG_GETARG_INT32(0);
+	backend_stats = pgstat_fetch_stat_backend_by_pid(pid, NULL);
+
+	if (!backend_stats)
+		PG_RETURN_NULL();
+
+	aio_counters = backend_stats->aio_counters;
+
+	/* Initialise attributes information in the tuple descriptor */
+	tupdesc = CreateTemplateTupleDesc(PG_STAT_BACKEND_AIO_COLS);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 1, "started",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 2, "executed_sync",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 3, "executed_async",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 4, "completed_self",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 5, "completed_other",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 6, "handle_waits",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 7, "submitted",
+					   INT8OID, -1, 0);
+	TupleDescInitEntry(tupdesc, (AttrNumber) 8, "stats_reset",
+					   TIMESTAMPTZOID, -1, 0);
+	TupleDescFinalize(tupdesc);
+	BlessTupleDesc(tupdesc);
+
+	/* Fill values */
+	values[0] = Int64GetDatum(aio_counters.started);
+	values[1] = Int64GetDatum(aio_counters.executed_sync);
+	values[2] = Int64GetDatum(aio_counters.executed_async);
+	values[3] = Int64GetDatum(aio_counters.completed_self);
+	values[4] = Int64GetDatum(aio_counters.completed_other);
+	values[5] = Int64GetDatum(aio_counters.handle_waits);
+	values[6] = Int64GetDatum(aio_counters.submitted);
+
+	if (backend_stats->stat_reset_timestamp != 0)
+		values[7] = TimestampTzGetDatum(backend_stats->stat_reset_timestamp);
+	else
+		nulls[7] = true;
+
+	PG_RETURN_DATUM(HeapTupleGetDatum(heap_form_tuple(tupdesc, values, nulls)));
+}
+
 /*
  * Returns statistics of WAL activity
  */
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 3cb84359adf..a089860ada4 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -6108,6 +6108,13 @@
   proargmodes => '{i,o,o,o,o,o}',
   proargnames => '{backend_pid,locktype,waits,wait_time,fastpath_exceeded,stats_reset}',
   prosrc => 'pg_stat_get_backend_lock' },
+{ oid => '9082', descr => 'statistics: backend AIO activity',
+  proname => 'pg_stat_get_backend_aio', provolatile => 'v', proparallel => 'r',
+  prorettype => 'record', proargtypes => 'int4',
+  proallargtypes => '{int4,int8,int8,int8,int8,int8,int8,int8,timestamptz}',
+  proargmodes => '{i,o,o,o,o,o,o,o,o}',
+  proargnames => '{backend_pid,started,executed_sync,executed_async,completed_self,completed_other,handle_waits,submitted,stats_reset}',
+  prosrc => 'pg_stat_get_backend_aio' },
 { oid => '6248', descr => 'statistics: information about WAL prefetching',
   proname => 'pg_stat_get_recovery_prefetch', prorows => '1', proretset => 't',
   provolatile => 'v', prorettype => 'record', proargtypes => '',
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 58a44857f13..64afb2bc082 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -514,6 +514,21 @@ typedef struct PgStat_WalStats
 	TimestampTz stat_reset_timestamp;
 } PgStat_WalStats;
 
+/* -------
+ * PgStat_AioCounters	AIO activity counters
+ * -------
+ */
+typedef struct PgStat_AioCounters
+{
+	PgStat_Counter started;
+	PgStat_Counter executed_sync;
+	PgStat_Counter executed_async;
+	PgStat_Counter completed_self;
+	PgStat_Counter completed_other;
+	PgStat_Counter handle_waits;
+	PgStat_Counter submitted;
+} PgStat_AioCounters;
+
 /* -------
  * PgStat_Backend		Backend statistics
  * -------
@@ -524,6 +539,7 @@ typedef struct PgStat_Backend
 	PgStat_BktypeIO io_stats;
 	PgStat_WalCounters wal_counters;
 	PgStat_PendingLock lock_stats;
+	PgStat_AioCounters aio_counters;
 } PgStat_Backend;
 
 /* ---------
@@ -542,6 +558,8 @@ typedef struct PgStat_BackendPending
 	 * PGSTAT_KIND_LOCK.
 	 */
 	PgStat_PendingLock pending_lock;
+	/* Store the AIO statistics counters */
+	PgStat_AioCounters aio_counters;
 } PgStat_BackendPending;
 
 /*
@@ -598,6 +616,13 @@ extern void pgstat_count_backend_io_op(IOObject io_object,
 extern void pgstat_count_backend_lock_waits(uint8 locktag_type, PgStat_Counter usecs);
 extern void pgstat_count_backend_lock_fastpath_exceeded(uint8 locktag_type);
 
+/* used by aio.c for AIO stats tracked in backends */
+extern void pgstat_count_backend_aio_start(bool synchronous);
+extern void pgstat_count_backend_aio_complete_self(void);
+extern void pgstat_count_backend_aio_complete_other(void);
+extern void pgstat_count_backend_aio_handle_wait(void);
+extern void pgstat_count_backend_aio_submitted(void);
+
 extern PgStat_Backend *pgstat_fetch_stat_backend(ProcNumber procNumber);
 extern PgStat_Backend *pgstat_fetch_stat_backend_by_pid(int pid,
 														BackendType *bktype);
diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h
index b3dc3ff7d8b..b2092fb42b9 100644
--- a/src/include/utils/pgstat_internal.h
+++ b/src/include/utils/pgstat_internal.h
@@ -706,7 +706,8 @@ extern void pgstat_archiver_snapshot_cb(void);
 #define PGSTAT_BACKEND_FLUSH_IO		(1 << 0)	/* Flush I/O statistics */
 #define PGSTAT_BACKEND_FLUSH_WAL   (1 << 1) /* Flush WAL statistics */
 #define PGSTAT_BACKEND_FLUSH_LOCK  (1 << 2) /* Flush lock statistics */
-#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK)
+#define PGSTAT_BACKEND_FLUSH_AIO   (1 << 3) /* Flush AIO statistics */
+#define PGSTAT_BACKEND_FLUSH_ALL   (PGSTAT_BACKEND_FLUSH_IO | PGSTAT_BACKEND_FLUSH_WAL | PGSTAT_BACKEND_FLUSH_LOCK | PGSTAT_BACKEND_FLUSH_AIO)
 
 extern bool pgstat_flush_backend(bool nowait, uint32 flags);
 extern bool pgstat_backend_flush_cb(bool nowait);
diff --git a/src/test/modules/test_aio/t/001_aio.pl b/src/test/modules/test_aio/t/001_aio.pl
index 63cadd64c15..8ecf3d3ad91 100644
--- a/src/test/modules/test_aio/t/001_aio.pl
+++ b/src/test/modules/test_aio/t/001_aio.pl
@@ -1842,6 +1842,37 @@ read_buffers('$table', 0, 4)|,
 	$psql_c->quit();
 }
 
+# Test per-backend AIO statistics counters
+sub test_aio_stats
+{
+	my $io_method = shift;
+	my $node = shift;
+
+	my $psql = $node->background_psql('postgres', on_error_stop => 0);
+
+	# Reset backend stats, evict relation, then read it back to force
+	# physical IO through the AIO layer.
+	$psql->query_safe(qq(SELECT pg_stat_reset_backend_stats(pg_backend_pid())));
+	$psql->query_safe(qq(SELECT evict_rel('tbl_ok')));
+	$psql->query_safe(qq(SELECT count(*) FROM tbl_ok));
+	$psql->query_safe(qq(SELECT pg_stat_force_next_flush()));
+
+	# started must be > 0 after physical IO
+	my $started = $psql->query_safe(
+		qq(SELECT started FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	cmp_ok($started, '>', 0,
+		"$io_method: AIO stats: started > 0 after physical IO");
+
+	# invariant: started = executed_sync + executed_async
+	my $consistent = $psql->query_safe(
+		qq(SELECT started = executed_sync + executed_async
+		   FROM pg_stat_get_backend_aio(pg_backend_pid())));
+	is($consistent, 't',
+		"$io_method: AIO stats: started = executed_sync + executed_async");
+
+	$psql->quit();
+}
+
 # Run all tests that for the specified node / io_method
 sub test_io_method
 {
@@ -1878,6 +1909,7 @@ CHECKPOINT;
 	test_ignore_checksum($io_method, $node);
 	test_checksum_createdb($io_method, $node);
 	test_read_buffers($io_method, $node);
+	test_aio_stats($io_method, $node);
 
 	# generic injection tests
   SKIP:
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 117e7379f10..6ce3077fa5b 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -2329,6 +2329,7 @@ PgStatShared_ReplSlot
 PgStatShared_SLRU
 PgStatShared_Subscription
 PgStatShared_Wal
+PgStat_AioCounters
 PgStat_ArchiverStats
 PgStat_Backend
 PgStat_BackendPending
-- 
2.34.1


--Pi0/Nu3LVhi0Ij+H--





^ permalink  raw  reply  [nested|flat] 201+ messages in thread