Andres Freund wrote:
> On 2014-11-10 14:28:30 -0300, Alvaro Herrera wrote:
> > If what we want is to quantify the extent of the issue, would it be more
> > convenient to save counters to pgstat? Vacuum already sends pgstat
> > messages, so there's no additional traffic there.
>
> I'm pretty strongly against that one in isolation. They'd need to be
> stored somewhere and they'd need to be queryable somewhere with enough
> context to make sense. To actually make sense of the numbers we'd also
> need to report all the other datapoints of vacuum in some form. That's
> quite a worthwile project imo - but *much* *much* more work than this.
We already have "last_autovacuum" columns and such in pg_stat_tables et
al, which only record the last value. My thinking regarding such
numbers is that you would save histories and put them in a chart, see
how they evolve with time. I doubt the individual numbers are worth
much, but the trends might show something interesting. As far as I
know, this is already true for most other pgstat values, with exception
of things such as live_tuples which are absolute numbers rather than
running counters.
I agree having more vacuuming data in general is a worthwhile project,
much larger than this one. Wasn't Greg Smith working on that?
--
Álvaro Herrera http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services