Re: SLRU statistics
Tomas Vondra <tomas.vondra@2ndquadrant.com>
From: Tomas Vondra <tomas.vondra@2ndquadrant.com>
To: Alvaro Herrera <alvherre@2ndquadrant.com>
Cc: Masahiko Sawada <masahiko.sawada@2ndquadrant.com>, "tsunakawa.takay@fujitsu.com" <tsunakawa.takay@fujitsu.com>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2020-03-28T00:26:30Z
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 →
-
Fix async.c to not register any SLRU stats counts in the postmaster.
- 7fd89f4d7a51 13.0 landed
-
Use proper GetDatum function in pg_stat_get_slru().
- 043e3e040160 13.0 landed
-
Initialize SLRU stats entries to zero.
- f3d23d83ef9a 13.0 landed
-
Remove pg_xact from pg_stat_reset_slru docs
- e685ca63ca4f 13.0 landed
-
Remove pg_xact entry from SLRU stats
- 2e08d314ed07 13.0 landed
-
Track SLRU page hits in SimpleLruReadPage_ReadOnly
- 2b88fdde30d8 13.0 landed
-
Fix typo in SLRU stats documentation
- 2c220ca46f3f 13.0 landed
-
Collect statistics about SLRU caches
- 28cac71bd368 13.0 landed
-
Simplify and rename some GUC variables, per various recent discussions:
- 48f7e6439568 8.3.0 cited
Attachments
- 0001-Collect-SLRU-statistics.patch (text/plain)
Hi,
here is a bit improved version of the patch - I've been annoyed by how
the resetting works (per-entry timestamp, but resetting all entries) so
I've added a new function pg_stat_reset_slru() that allows resetting
either all entries or just one entry (identified by name). So
SELECT pg_stat_reset_slru('clog');
resets just "clog" SLRU counters, while
SELECT pg_stat_reset_slru(NULL);
resets all entries.
I've also done a bit of benchmarking, to see if this has measurable
impact (in which case it might deserve a new GUC), and I think it's not
measurable. I've used a tiny unlogged table (single row).
CREATE UNLOGGED TABLE t (a int);
INSERT INTO t VALUES (1);
and then short pgbench runs with a single client, updatint the row. I've
been unable to measure any regression, it's all well within 1% so noise.
But perhaps there's some other benchmark that I should do?
regards
--
Tomas Vondra http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services