Re: postgres query log analysis? - Mailing list pgsql-general

From Larry Rosenman
Subject Re: postgres query log analysis?
Date
Msg-id 02d501c6ebcf$e61a51c0$0202fea9@SURGIENT.COM
Whole thread Raw
In response to postgres query log analysis?  ("George Pavlov" <gpavlov@mynewplace.com>)
Responses Re: postgres query log analysis?  ("George Pavlov" <gpavlov@mynewplace.com>)
List pgsql-general
George Pavlov wrote:
> What do you all use for query log analysis for Postgres. I feel/hope
> like there must be something that I am missing.
>
> I have tried PQA (http://pqa.projects.postgresql.org/) and it is very
> problematic, at least with the kind of application we have. Some of
>   the problems: * not aware of prepared statements
> (http://pgfoundry.org/forum/forum.php?thread_id=948&forum_id=24);
>   * giving faulty stats in several situations (possibly related to
> prepared statements, but maybe some others);
>   * very slow on anything but "play" logs;
>   * seems to not be maintained lately (since 8.0?).
>
> Pglog-analyze (http://opensource.pearshealthcyber.cz/) is an
> admirable attempt, but it is very problematic too:
>   * incredibly memory-intensive -- I can at best process a few hours
> of my production logs even though I have jacked up my PHP memory
> limit to 2.5GB
>   * gives out stats that seem way off on several accounts (subsecond
> queries sometimes show as taking over a minute);
>   * seems to have similar to PQA's (if not worse) issues with
> prepared statements.
>
> Is there any other option? When you enable query logging in your
> PGSQL logs what do you do with the output?
>
> Thanks!
>


Look into pgfouine on pgFoundry. http://pgfoundry.org/projects/pgfouine/


--
Larry Rosenman                     http://www.lerctr.org/~ler
Phone: +1 512-248-2683             E-Mail: ler@lerctr.org
US Mail: 430 Valona Loop, Round Rock, TX 78681-3893


pgsql-general by date:

Previous
From: "George Pavlov"
Date:
Subject: postgres query log analysis?
Next
From: "Lenorovitz, Joel"
Date:
Subject: Determining caller of a function (due to a cascaded FK constraint?)