Re: RFC: Timing Events - Mailing list pgsql-hackers

From Albe Laurenz
Subject Re: RFC: Timing Events
Date
Msg-id D960CB61B694CF459DCFB4B0128514C208A4E909@exadv11.host.magwien.gv.at
Whole thread Raw
In response to Re: RFC: Timing Events  (Josh Berkus <josh@agliodbs.com>)
Responses Re: RFC: Timing Events  (Josh Berkus <josh@agliodbs.com>)
List pgsql-hackers
Josh Berkus wrote:
>> Huh?  The typical use-case is to enable it for all sessions by
>> including it in shared_preload_libraries.  That doesn't require any
>> particular session to be superuser.  (If you're superuser you can then
>> turn it *off* in your session, should you wish.)
> 
> It's not practical to have auto-explain on for all queries on a server
> which is processing 10K queries/minute.  And non-superusers can't alter
> the settings in their session, even the min_duration.
> 
> A practical use of auto-explain would involve during it on for a single
> user session, or for a specific database user, neither of which are
> possible.

I think that Tom is right.

You can reduce the logging volume on busy servers with
auto_explain.log_min_duration.  You can also activate it for a single
database user only by setting log_min_duration to -1 globally and
change the setting for one user with ALTER ROLE SET, right?

And I am under the impression that parameters that influence
what is logged tend to be superuser only for a good reason
(the only exception I see are the debug_print_* parameters).

Yours,
Laurenz Albe

pgsql-hackers by date:

Previous
From: Simon Riggs
Date:
Subject: Re: Arguments to foreign tables?
Next
From: Shigeru HANADA
Date:
Subject: Re: FDW for PostgreSQL