Comments on: Human Factors Behind Incidents: Why Settings Like “idle_session_timeout” Can Be a Bad Idea https://www.percona.com/blog/human-factors-behind-incidents-why-settings-like-idle_session_timeout-can-be-a-bad-idea/ Thu, 08 Feb 2024 21:50:55 +0000 hourly 1 https://wordpress.org/?v=6.5.2 By: WEKSLEY MESQUITA VIANA https://www.percona.com/blog/human-factors-behind-incidents-why-settings-like-idle_session_timeout-can-be-a-bad-idea/#comment-10974264 Thu, 08 Feb 2024 21:50:55 +0000 https://www.percona.com/blog/?p=92930#comment-10974264 Firstly congratulations for the article. Its very enlightening about the use of the idle_session_timeout parameter.

I wanted to ask a question about the reference where it says: “The max_connections setting of a PostgreSQL instance should not be too far from 10x of available CPUs.”

I’m looking for recent references that help define this max_connections when using SSDs. The literature only deals with old HDDs that point to something between 2 and 4x the number of processors.

I found this 10x number very interesting.

]]>