Re: compute_query_id and pg_stat_statements

On Sat, May 15, 2021 at 04:09:32PM +0800, Julien Rouhaud wrote:
> > While thinking about this patch it occurred to that an useful gadget
> > might be to let pg_stat_statements be loaded always, but with
> > compute_query_id=false; so it's never active; except if you
> >   ALTER {DATABASE, USER} foo SET (compute_query_id=on);
> > so that it's possible to enable it selectively.  I think this doesn't
> > currently achieve anything because pgss_store is always called
> > regardless of query ID being active (so you'd always have at least one
> > function call as performance penalty, only that the work would be for
> > naught), but that seems a simple change to make.  I didn't look closely
> > to see what other things would need patched.
> 
> Couldn't it already be achieved with ALTER [ DATABASE | USER ] foo SET
> pg_stat_statements.track = [ none | top | all ] ?

I am no longer the committer in charge of this feature, but I would like
to remind the group that beta1 will be wrapped on Monday, and it is hard
to change non-read-only GUCs after beta since the settings are embedded
in postgresql.conf.  There is also a release notes item that probably
will need to be adjusted.

-- 
  Bruce Momjian  <bruce@momjian.us>        https://momjian.us
  EDB                                      https://enterprisedb.com

  If only the physical world exists, free will is an illusion.