pg.ddx.io pgsql-admin@postgresql.org mailing list archive
help / color / mirror / Atom feedMonitoring autovacuum induced contention
2+ messages / 2 participants
[nested] [flat]
* Monitoring autovacuum induced contention
@ 2023-12-06 21:35 Joseph Hammerman <joe.hammerman@datadoghq.com>
0 siblings, 1 reply; 2+ messages in thread
From: Joseph Hammerman @ 2023-12-06 21:35 UTC (permalink / raw)
To: pgsql-admin
Hi admins,
We are working on tuning autovacuum correctly. While the biggest problem we
have and most other OLTP deployments I have seen have is autovacuum
resource starvation, I'd like to be able to reliably determine when we have
turned autovacuum capacity dials too far.
Is there any way to track autovacuum induced query waits?
Thanks,
Joe
^ permalink raw reply [nested|flat] 2+ messages in thread
* Re: Monitoring autovacuum induced contention
@ 2023-12-07 03:16 Laurenz Albe <laurenz.albe@cybertec.at>
parent: Joseph Hammerman <joe.hammerman@datadoghq.com>
0 siblings, 0 replies; 2+ messages in thread
From: Laurenz Albe @ 2023-12-07 03:16 UTC (permalink / raw)
To: Joseph Hammerman <joe.hammerman@datadoghq.com>; pgsql-admin
On Wed, 2023-12-06 at 13:35 -0800, Joseph Hammerman wrote:
> We are working on tuning autovacuum correctly. While the biggest problem we have
> and most other OLTP deployments I have seen have is autovacuum resource starvation,
> I'd like to be able to reliably determine when we have turned autovacuum capacity
> dials too far.
>
> Is there any way to track autovacuum induced query waits?
Autovacuum won't directly block you, so you will never see a process waiting for
autovacuum. The exception to this rule is that statements that take high locks,
like CREATE INDEX or TRUNCATE, may be seen waiting for autovacuum to complete
occasionally. But that is probably not your problem.
Autovacuum will compete for system resources with your normal workload.
So you should monitor operating system resources like CPU or disk I/O, and if
any of these is at capacity while autovacuum is processing a table, you will
feel the impact.
Yours,
Laurenz Albe
^ permalink raw reply [nested|flat] 2+ messages in thread
end of thread, other threads:[~2023-12-07 03:16 UTC | newest]
Thread overview: 2+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2023-12-06 21:35 Monitoring autovacuum induced contention Joseph Hammerman <joe.hammerman@datadoghq.com>
2023-12-07 03:16 ` Laurenz Albe <laurenz.albe@cybertec.at>
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox