Thread

Re: Enhance statistics reset functions to return reset timestamp

Shinya Kato <shinya11.kato@gmail.com> — 2025-09-12T06:31:52Z

Hi,

On Wed, Aug 13, 2025 at 4:06 PM Shinya Kato <shinya11.kato@gmail.com> wrote:
>
> On Sat, Aug 9, 2025 at 7:37 AM Andres Freund <andres@anarazel.de> wrote:
> >
> > Hi,
> >
> > On 2025-08-08 13:18:39 +0900, Shinya Kato wrote:
> > > I would like to propose a series of patches that enhance the behavior
> > > of various statistics reset functions by making them return the reset
> > > timestamp. This change improves usability and aligns with the behavior
> > > introduced in commit dc9f8a798[1], where pg_stat_statements_reset()
> > > was updated to return the reset time.
> > >
> > > The following functions have been modified to return a TIMESTAMP WITH
> > > TIME ZONE value indicating when the statistics were reset:
> > > - pg_stat_reset()
> > > - pg_stat_reset_shared()
> > > - pg_stat_reset_single_table_counters()
> > > - pg_stat_reset_backend_stats()
> > > - pg_stat_reset_single_function_counters()
> > > - pg_stat_reset_slru()
> > > - pg_stat_reset_replication_slot()
> > > - pg_stat_reset_subscription_stats()
> > > - pg_stat_clear_snapshot()
> >
> > -1 - I think it was a mistake to introduce support for granular resets, we
> > shouldn't bury ourselves deeper.  If anything we should rip out everything
> > other than 1) a global reset b) a per-database reset.
> >
> > Leaving that aside, I just don't see a convincing use case for returning the
> > timestamp here.
>
> As I mentioned earlier, the use case is to obtain an audit trail of
> the statistics reset time from the function’s return value.
>
> Also, how should we think about consistency with
> pg_stat_statements_reset()? The reset timestamp for pg_stat_statements
> can be obtained from the pg_stat_statements_info view, which returns a
> TIMESTAMPTZ. Of course, pg_stat_statements is a contrib module, so we
> don’t necessarily have to match its behavior. However, when running a
> script that resets various statistics before tests, if only
> pg_stat_statements_reset() returns a TIMESTAMPTZ, it might look as
> though the other statistics reset functions failed.
>
> What these patches do is simply return the TIMESTAMPTZ that is already
> computed for the pg_stat_* views, which seems to me a reasonable
> change.

Rebased the patches.

-- 
Best regards,
Shinya Kato
NTT OSS Center