Content-Type: multipart/mixed; boundary="------------5A1EOYBs1isLWH0SMAZDnoGA"
Message-ID: <3546e2a7-5831-4c0a-996b-358d4ec483a6@tantorlabs.com>
Date: Mon, 10 Feb 2025 11:05:29 +0300
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: Sample rate added to pg_stat_statements
From: Ilia Evdokimov <ilya.evdokimov@tantorlabs.com>
To: Sami Imseih <samimseih@gmail.com>,
 Alexander Korotkov <aekorotkov@gmail.com>,
 Michael Paquier <michael@paquier.xyz>,
 Greg Sabino Mullane <htamfids@gmail.com>,
 Andrey Borodin <x4mmm@yandex-team.ru>,
 Alena Rybakina <a.rybakina@postgrespro.ru>
Cc: pgsql-hackers <pgsql-hackers@postgresql.org>
References: <fe99e0ca-e564-480e-b865-5f0cee30bc60@tantorlabs.com>
 <6cf6270b-e8fa-4d99-b74a-17e781cdf309@tantorlabs.com>
 <b40b76c3-fc54-4f11-a50f-e63cd6434560@postgrespro.ru>
 <3a27c830-b81f-4560-9774-a87d9b922fbd@tantorlabs.com>
 <CAA5RZ0vYu+dm3KKd5iaQ=LY0J2sjfxJG+-VCfkeVtWFuhSb-4A@mail.gmail.com>
 <cebfee9c-a894-46f1-b6bc-5bf4a0753b41@tantorlabs.com>
 <7AD765EF-14DE-4B3E-8282-E82A9B44DBAA@yandex-team.ru>
 <e738eac5-ad4d-454c-a079-38766e32ce52@tantorlabs.com>
 <CAA5RZ0s+khQ_u84qS_cAHzxVQQHK_UmWXxRBWTTex7zW1250tQ@mail.gmail.com>
 <0c7770bd-9750-4200-a472-bdb262583c2f@tantorlabs.com>
 <CAA5RZ0vxn_UiUkUK5SdngObLZzw40RhafB7SHydxzBw2_xjjiA@mail.gmail.com>
 <cc1beb52-3e4e-4866-8c32-a5967b98c977@tantorlabs.com>
 <18631d46-1741-4edc-b116-8d9631cdf919@tantorlabs.com>
 <4f98ec33-c627-4eb4-84f3-9994a0982277@tantorlabs.com>
 <CAA5RZ0tJjrX6YiQVDZCOA_BDMuhO-po_T8a3PmjxEEH_iL-qMg@mail.gmail.com>
 <623db69f-b40a-4d50-b1bd-bb073fcc4336@tantorlabs.com>
Content-Language: en-US
In-Reply-To: <623db69f-b40a-4d50-b1bd-bb073fcc4336@tantorlabs.com>
Archived-At: <https://www.postgresql.org/message-id/3546e2a7-5831-4c0a-996b-358d4ec483a6%40tantorlabs.com>
Precedence: bulk

This is a multi-part message in MIME format.
--------------5A1EOYBs1isLWH0SMAZDnoGA
Content-Type: multipart/alternative;
 boundary="------------lzyosQTpEPnz8nyRjErIjQn9"

--------------lzyosQTpEPnz8nyRjErIjQn9
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit

Hi hackers,

Since current patch is in the commitfest with the status 'Ready for 
committer', I’d like to summarize what it does, the problems it 
addresses, and answer the key questions raised in the discussion thread.

Enabling pg_stat_statements can cause a performance drop due to two main 
reasons:

  * Contention on the exclusive lock, which blocks the hash table when
    allocating or deallocating an entry.
  * Contention on the spinlock, which blocks access to an existing entry.

The exclusive lock issue can be mitigated by normalizing similar queries 
to the same queryid. There is an ongoing thread discussing this approach 
for IN and ARRAY queries: [0]. Another solution is integrating 
pg_stat_statements into central pgstats with the custom APIs, without 
pushing the module into core [1].

However, even if we do all that, the spinlock contention problem will 
persist for very frequent queries. This is exactly what I aim to solve 
with sampling. Other approaches, such as replacing the spinlock with 
atomic variables or an lwlock, have not yielded good results [2]. Even 
track_planning was disabled by default to reduce contention, but on 
high-core systems, the issue still remains.So far, I see no alternative 
solution to this problem.

The benchmark results on different number of CPUs demonstrating the 
impact of this patch can be found here:

pgbench -c{CPUs} -j20 -S -Mprepared -T120
   CPUs | sample_rate |   tps | CPU waits | ClientRead wait | SpinDelay wait
    192 |         1.0 | 484338|      9568 |             929 | 11107
    192 |        0.75 | 909547|     12079 |            2100 | 4781
    192 |         0.5 |1028594|     13253 |            3378 | 174
    192 |        0.25 |1019507|     13397 |            3423 | -
    192 |         0.0 |1015425|     13106 |            3502 | -

     48 |         1.0 | 643251|       911 |             932 | 44
     48 |        0.75 | 653946|       939 |             939 |  3
     48 |         0.5 | 651654|       841 |             932 | -
     48 |        0.25 | 652668|       860 |             910 | -
     48 |         0.0 | 659111|       849 |             882 | -

     32 |         1.0 | 620667|      1782 |             560 | -
     32 |        0.75 | 620663|      1736 |             554 | -
     32 |         0.5 | 624094|      1741 |             648 | -
     32 |        0.25 | 628638|      1702 |             576 | -
     32 |         0.0 | 630483|      1638 |             574 | -

Some suggested sampling based on execution time instead of a random 
generator. While this is a good idea, it has a critical limitation. 
Execution time can only be checked in pgss_ExecutorEnd. However, 
pgss_planner already interacts with the spinlock, and we can not sample 
based on execution time at that stage. Finding a criterion that works 
for both pgss_planner and pgss_ExecutorEnd is difficult. This is why 
random sampling remains the most viable option.

BTW, Instead of using a bool flag to check if we are inside 
pgss_post_parse_analyze, we now pass a pgssStoreKind to pgss_enabled. 
This makes the logic clearer and more readable, explicitly indicating 
where we need to check whether a query should be sampled. I made it in 
new v15 patch.

Are there any remaining concerns or alternative suggestions regarding 
this approach? I'm happy to discuss further.

[0]: https://commitfest.postgresql.org/51/2837/
[1]: https://www.postgresql.org/message-id/Zz0MFPq1s4WFxxpL%40paquier.xyz
[2]: https://commitfest.postgresql.org/29/2634/

--
Best regards,
Ilia Evdokimov,
Tantor Labs LLC.

--------------lzyosQTpEPnz8nyRjErIjQn9
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: 8bit

<!DOCTYPE html>
<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <p>Hi hackers,</p>
    <p>Since current patch is in the commitfest with the status 'Ready
      for committer', I’d like to summarize what it does, the problems
      it addresses, and answer the key questions raised in the
      discussion thread.</p>
    <p>Enabling pg_stat_statements can cause a performance drop due to
      two main reasons:</p>
    <ul>
      <li>Contention on the exclusive lock, which blocks the hash table
        when allocating or deallocating an entry.</li>
      <li>Contention on the spinlock, which blocks access to an existing
        entry.</li>
    </ul>
    <p>The exclusive lock issue can be mitigated by normalizing similar
      queries to the same queryid. There is an ongoing thread discussing
      this approach for IN and ARRAY queries: [0]. Another solution is
      integrating pg_stat_statements into central pgstats with the
      custom APIs, without pushing the module into core [1].</p>
    <p>However, even if we do all that, the spinlock contention problem
      will persist for very frequent queries. This is exactly what I aim
      to solve with sampling. Other approaches, such as replacing the
      spinlock with atomic variables or an lwlock, have not yielded good
      results [2]. Even track_planning was disabled by default to reduce
      contention, but on high-core systems, the issue still remains.So
      far, I see no alternative solution to this problem.</p>
    <p>The benchmark results on different number of CPUs demonstrating
      the impact of this patch can be found here:</p>
    <p>pgbench -c{CPUs} -j20 -S -Mprepared -T120<br>
        CPUs | sample_rate |   tps | CPU waits | ClientRead wait |
      SpinDelay wait<br>
         192 |         1.0 | 484338|      9568 |             929 | 11107<br>
         192 |        0.75 | 909547|     12079 |            2100 | 4781<br>
         192 |         0.5 |1028594|     13253 |            3378 | 174<br>
         192 |        0.25 |1019507|     13397 |            3423 | -<br>
         192 |         0.0 |1015425|     13106 |            3502 | -<br>
      <br>
          48 |         1.0 | 643251|       911 |             932 | 44<br>
          48 |        0.75 | 653946|       939 |             939 |  3<br>
          48 |         0.5 | 651654|       841 |             932 | -<br>
          48 |        0.25 | 652668|       860 |             910 | -<br>
          48 |         0.0 | 659111|       849 |             882 | -<br>
      <br>
          32 |         1.0 | 620667|      1782 |             560 | -<br>
          32 |        0.75 | 620663|      1736 |             554 | -<br>
          32 |         0.5 | 624094|      1741 |             648 | -<br>
          32 |        0.25 | 628638|      1702 |             576 | -<br>
          32 |         0.0 | 630483|      1638 |             574 | -<br>
    </p>
    <p>Some suggested sampling based on execution time instead of a
      random generator. While this is a good idea, it has a critical
      limitation. Execution time can only be checked in
      pgss_ExecutorEnd. However, pgss_planner already interacts with the
      spinlock, and we can not sample based on execution time at that
      stage. Finding a criterion that works for both pgss_planner and
      pgss_ExecutorEnd is difficult. This is why random sampling remains
      the most viable option.</p>
    <p>BTW, Instead of using a bool flag to check if we are inside
      pgss_post_parse_analyze, we now pass a pgssStoreKind to
      pgss_enabled. This makes the logic clearer and more readable,
      explicitly indicating where we need to check whether a query
      should be sampled. I made it in new v15 patch.</p>
    <p>Are there any remaining concerns or alternative suggestions
      regarding this approach? I'm happy to discuss further.</p>
    <p>[0]: <a class="moz-txt-link-freetext" href="https://commitfest.postgresql.org/51/2837/">https://commitfest.postgresql.org/51/2837/</a> <br>
      [1]:
      <a class="moz-txt-link-freetext" href="https://www.postgresql.org/message-id/Zz0MFPq1s4WFxxpL%40paquier.xyz">https://www.postgresql.org/message-id/Zz0MFPq1s4WFxxpL%40paquier.xyz</a><br>
      [2]: <a class="moz-txt-link-freetext" href="https://commitfest.postgresql.org/29/2634/">https://commitfest.postgresql.org/29/2634/</a></p>
    <p>--<br>
      Best regards,<br>
      Ilia Evdokimov,<br>
      Tantor Labs LLC.<br>
    </p>
  </body>
</html>

--------------lzyosQTpEPnz8nyRjErIjQn9--
--------------5A1EOYBs1isLWH0SMAZDnoGA
Content-Type: text/x-patch; charset=UTF-8;
 name="v15-0001-Allow-setting-sample-rate-for-pg_stat_statements.patch"
Content-Disposition: attachment;
 filename*0="v15-0001-Allow-setting-sample-rate-for-pg_stat_statements.pa";
 filename*1="tch"
Content-Transfer-Encoding: base64
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--------------5A1EOYBs1isLWH0SMAZDnoGA--