From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Sayyid Ali Sajjad Rizavi <sasrizavi(at)gmail(dot)com> |
Cc: | pgsql-hackers(at)lists(dot)postgresql(dot)org |
Subject: | Re: Enable pg_stat_statements extension for limited statements only |
Date: | 2022-11-30 18:15:04 |
Message-ID: | 1434937.1669832104@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Sayyid Ali Sajjad Rizavi <sasrizavi(at)gmail(dot)com> writes:
> Hi, I'd like to propose a change and get advice if I should work on it.
> The extension pg_stat_statements is very helpful, but the downside is that
> it will take up too much disk space when storing query stats if it's
> enabled for all statements like SELECT, INSERT, UPDATE, DELETE.
It will only take up a lot of disk space if you let it, by setting
the pg_stat_statements.max parameter too high.
> For example, deletes do not happen too frequently; so I'd like to be able
> to enable pg_stat_statements only for the DELETE statement, maybe using
> some flags.
I'm a little skeptical of the value of that. Why would you want stats
only for infrequent statements?
I'm not denying that there might be usefulness in filtering what
pg_stat_statements will track, but it's not clear to me that
this particular proposal will be useful to many people.
I wonder whether there would be more use in filters expressed
as regular expressions to match against the statement text.
That would allow, for example, tracking statements that mention
a particular table as well as statements with a particular
head keyword. I could see usefulness in both a positive filter
(must match this to get tracked) and a negative one (must not
match this to get tracked).
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Daniel Verite | 2022-11-30 18:22:42 | Re: Tests for psql \g and \o |
Previous Message | Alvaro Herrera | 2022-11-30 18:12:01 | Re: generic plans and "initial" pruning |