Re: Adding wait events statistics

Hi,

On Fri, Jul 11, 2025 at 11:59:39AM -0400, Andres Freund wrote:
> Hi,
>
> On 2025-07-10 08:21:56 +0000, Bertrand Drouvot wrote:
> > - Let's define a list of "cheap" and a list of "expensive" wait events
> > - Change the code path to increment counters only for the "expensive" ones
> > - Optimize the counter incrementation as much as possible
> >
> > Does that sound reasonable to you?
>
> That does seem like the minimum.
>
> Unfortunately I'm rather doubtful this provides enough value to be worth the
> cost. But I'm rather willing to be proven wrong.
>

I guess that would (among other things) depend of how many wait events would be
tracked. I'll do some study to measure the counters's impact on the wait events.

BTW, I started to do so by using (i.e adding) probes in pgstat_report_wait_start()
and pgstat_report_wait_end() (to produce something like the attached), and
I wonder if those probes could be added in the Built-in ones, what do you
think? (that's true that they could be added manually as I did but I'm not sure
about what "deserve" to be added in the Built=in list)).

Regards,

--
Bertrand Drouvot
PostgreSQL Contributors Team
RDS Open Source Databases
Amazon Web Services: https://aws.amazon.com