On 2022-07-09 03:18, Peter Geoghegan wrote:
> On Fri, Jul 8, 2022 at 10:47 AM Alvaro Herrera
> <alvherre@alvh.no-ip.org> wrote:
>> Saving some sort of history would be much more useful, but of course a
>> lot more work.
Thank you for the comments!
Yes, having some sort of history would be ideal in this case.
However, I am not sure how to implement those features at this moment,
so I will take some time to consider.
At the same time, I think having this metrics exposed in the
pg_stat_all_tables comes in handy when tuning the
maintenance_work_mem/autovacuume_work_mem even though it shows the value
of only last vacuum/autovacuum.
Regards,
--
Ken Kato
Advanced Computing Technology Center
Research and Development Headquarters
NTT DATA CORPORATION
Regression is failing on all platforms; please correct that and resubmit the patch.
[06:17:08.194] Failed test: 2
[06:17:08.194] Non-zero exit status: 1
[06:17:08.194] Files=33, Tests=411, 167 wallclock secs ( 0.20 usr 0.05 sys + 37.96 cusr 21.61 csys = 59.82 CPU)
[06:17:08.194] Result: FAIL
[06:17:08.194] make[2]: *** [Makefile:23: check] Error 1
[06:17:08.194] make[1]: *** [Makefile:52: check-recovery-recurse] Error 2
[06:17:08.194] make: *** [GNUmakefile:71: check-world-src/test-recurse] Error 2
--