Horton

Re: Vacuum statistics

Jim Nasby <jnasby@upgrade.com>

View thread
From: Jim Nasby <jnasby@upgrade.com>
To: Sami Imseih <samimseih@gmail.com>
Cc: Alena Rybakina <a.rybakina@postgrespro.ru>, Ilia Evdokimov <ilya.evdokimov@tantorlabs.com>, Andrei Zubkov <zubkov@moonset.ru>, Masahiko Sawada <sawada.mshk@gmail.com>, Melanie Plageman <melanieplageman@gmail.com>, jian he <jian.universality@gmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>, a.lepikhov@postgrespro.ru, Alexander Korotkov <aekorotkov@gmail.com>
Date: 2025-01-03T19:08:02Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Move wal_buffers_full from PgStat_PendingWalStats to WalUsage

On Jan 2, 2025, at 4:33 PM, Sami Imseih <samimseih@gmail.com> wrote:
> 
>> While backwards compatibility is important, there’s definitely precedent for changing
>> what shows up in the catalog. IMHO it’s better to bite the bullet and move those fields
>> instead of having vacuum stats spread across two different views.
> 
> Correct, the most recent one that I could think of is pg_stat_checkpointer,
> which pulled the checkpoint related columns from pg_stat_bgwriter.
> In that case though, these are distinct background processes and
> it's a clear distinction.
> 
> In this case, I am not so sure about this, particularly because
> we will then have the autoanalyze and autovacuum fields in different
> views, which could be more confusing to users than saying pg_stat_all_tables
> has high level metrics about vacuum and analyze and for more details on
> vacuum, refer to pg_stat_vacuum_tables ( or whatever name we settle on ).

I guess one question is how realistic it is to try and put everything about (auto)vacuum in a single view. Given the complexity, the answer to that might just be “no”. In that case leaving existing fields in pg_stat_all_tables is a lot more reasonable.

Related to this… it’d be nice if we had a view that gave insight to users about auto vacuum scheduling. I know there’s one floating around the internet, but given the number of systems I’ve seen where autovac can’t keep up it’d be good to raise user awareness.