public inbox for [email protected]  

help / color / mirror / Atom feed

From: Sami Imseih <[email protected]>
To: Michael Paquier <[email protected]>
Cc: Bertrand Drouvot <[email protected]>
Cc: pgsql-hackers <[email protected]>
Subject: Re: POC: track vacuum/analyze cumulative time per relation
Date: Fri, 24 Jan 2025 14:19:59 -0600
Message-ID: <CAA5RZ0s9FSC8cU65D_XeV19-Zd3kFgJrBokGjq-=_Ccpjk6ZLw@mail.gmail.com> (raw)
In-Reply-To: <[email protected]>
References: <[email protected]>
	<CAA5RZ0tP6my7Mbr_syP7hRAiFANC_bG8SyoQF=-vfV9VawpV=g@mail.gmail.com>
	<[email protected]>
	<CAA5RZ0v9060p7hjtLPY9i0Me6SoaRKYwxATqt_DnZr8dGSnUhQ@mail.gmail.com>
	<Z4ez/[email protected]>
	<CAA5RZ0s8=wd-X4GgMcFQO1WbwPtw4HjTtMcDahsnegZ8p6tYPQ@mail.gmail.com>
	<[email protected]>
	<CAA5RZ0vKWKULGZ_V8DOOtFHXtsOLMVLCpkJ+guniB66T-o_eZA@mail.gmail.com>
	<[email protected]>
	<CAA5RZ0su2Unc+iJkhL6QEfsbjj6-WXjRWyf2E0b4BotUH=o4FA@mail.gmail.com>
	<[email protected]>

> I was referring to the order of the fields in the structure itself,
> but that's no big deal one way or the other.

I understand your point now. I will group them with the
related counters in the next rev and will use

> This should be one comment for the whole block, or this should use the
> singular for each comment.

I will use a singular "/* time in milliseconds */" comment for each
new field.

This existing write_time field is plural, but I will leave
that one alone for now.

PgStat_Counter write_time; /* times in milliseconds */


> On HEAD in the ANALYZE path, the endtime is calculated after
> index_vacuum_cleanup().  Your patch calculates it before
> index_vacuum_cleanup(), which would result in an incorrect calculation
> if the index cleanup takes a long time with less logs generated, no?
> Sorry for not noticing that earlier on the thread..  Perhaps it would
> be just better to pass the start time to pgstat_report_vacuum() and
> pgstat_report_analyze() then let their internals do the elapsed time
> calculations.  Consistency of the arguments for both functions is
> something worth having, IMO, even if it means a bit more
> GetCurrentTimestamp() in this case.

So currently, the report of the last_(autoanalyze|analyze)_time
is set before the index_vacuum_cleanup, but for logging purposes
the elapsed time is calculated afterwards. Most users will not notice
this, but I think that is wrong as well.

I think we should calculate endtime and elapsedtime and call
pgstat_report_analyze after the index_vacuum_cleanup; and
before vac_close_indexes. This is more accurate and will
avoid incurring the extra GetCurrentTimestamp() call.

What do you think?

Regards,

Sami

view thread (19+ messages)  latest in thread

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients using the --to and --cc options:
  reply via email

  To: [email protected]
  Cc: [email protected], [email protected], [email protected]
  Subject: Re: POC: track vacuum/analyze cumulative time per relation
  In-Reply-To: <CAA5RZ0s9FSC8cU65D_XeV19-Zd3kFgJrBokGjq-=_Ccpjk6ZLw@mail.gmail.com>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox