Re: survey: pg_stat_statements total_time and entry deallocation

From: legrand legrand <legrand_legrand(at)hotmail(dot)com>
To: pgsql-general(at)postgresql(dot)org
Subject: Re: survey: pg_stat_statements total_time and entry deallocation
Date: 2018-10-05 22:33:16
Message-ID: 1538778796328-0.post@n3.nabble.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Hello,

What about adding a log message for each entry_dealloc() execution?
it could be usefull to start thinking increasing pg_stat_statements.max.

is there any rule regarding the acceptable max value ? I'm playing in test
with a 20 000 value without any problem, could it extendend to 100 000 ?

In a system with pressure on numbers of pgss lines, and regular
entry_dealloc() executions, low frequency entries are evicted firsts, and
this should still be the same for new small queries with usage based on
total_time.

Maybe there is a third way, that would be to evict queries based on the
"oldest modification time" ...
This would garantee that latest queries would be kept long enough to be
collected by aggregation tools.

Regards
PAscal

--
Sent from: http://www.postgresql-archive.org/PostgreSQL-general-f1843780.html

In response to

Browse pgsql-general by date

  From Date Subject
Next Message ROS Didier 2018-10-06 09:57:25 Why the index is not used ?
Previous Message Jeremy Schneider 2018-10-05 20:47:25 Re: survey: pg_stat_statements total_time and entry deallocation