From: Rui DeSousa <rui.desousa@icloud.com>
To: Joseph Hammerman <joe.hammerman@datadoghq.com>
Cc: pgsql-admin@lists.postgresql.org
Subject: Re: Getting out ahead of OOM
Date: Fri, 7 Mar 2025 15:17:37 -0500
Message-ID: <F71FFCF6-19D1-49D2-8286-B71903373A9B@icloud.com> (raw)
In-Reply-To: <CAHs7QM_8SRfybxg_Pq3orGw5anMkV7OzKYfxDXSBQVWzcXf3Bw@mail.gmail.com>
References: <CAHs7QM_8SRfybxg_Pq3orGw5anMkV7OzKYfxDXSBQVWzcXf3Bw@mail.gmail.com>
> On Mar 7, 2025, at 2:07 PM, Joseph Hammerman <joe.hammerman@datadoghq.com> wrote:
>
> Has anyone had success tracking all the Postgres memory allocation configurables and using that to administratively prevent OOMing?
Don’t use memory limits in Kubernetes and we also run Postgres on dedicated Kubernetes clusters.
Shared memory will get counted multiple times. Each login session; as it maps in the shared buffers, it will get wrongly counted as memory used (it is shared memory!).
I have instances running on Kubernetes that only use 6GB of memory; however Kubernetes is wrongly reporting 50GB used due to number of active seasons. Our Postgres pods use to get terminated when “exceeding" the limit but not! Until removed the use of memory limits in Kubernetes.
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: pgsql-admin@postgresql.org
Cc: rui.desousa@icloud.com, joe.hammerman@datadoghq.com, pgsql-admin@lists.postgresql.org
Subject: Re: Getting out ahead of OOM
In-Reply-To: <F71FFCF6-19D1-49D2-8286-B71903373A9B@icloud.com>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox