Re: monitoring usage count distribution

On Tue, Apr 4, 2023 at 7:29 PM Andres Freund <andres(at)anarazel(dot)de> wrote:
> > I'm skeptical that pg_buffercache_summary() is a good idea at all
>
> Why? It's about two orders of magnitude faster than querying the equivalent
> data by aggregating in SQL. And knowing how many free and dirty buffers are
> over time is something quite useful to monitor / correlate with performance
> issues.

Well, OK, fair point.

> > but having it display the average usage count seems like a particularly poor
> > idea. That information is almost meaningless.
>
> I agree there are more meaningful ways to represent the data, but I don't
> agree that it's almost meaningless. It can give you a rough estimate of
> whether data in s_b is referenced or not.

I might have overstated my case.

> > Replacing that with a six-element integer array would be a clear improvement
> > and, IMHO, better than adding yet another function to the extension.
>
> I'd have no issue with that.

Cool.

--
Robert Haas
EDB: http://www.enterprisedb.com