Re: Max connections reached without max connections reached

[Adrian Klaver <adrian.klaver@xxxxxxxxxxx>]

On 11/23/21 16:58, James Sewell wrote:
>      > re: EnterpriseDB yes it is - I'm having this same discussion with
>     them
>      > in parallel
>
>     What version of theirs?
>
> PostgreSQL 11.9 (EnterpriseDB Advanced Server 11.9.18) on 
> x86_64-pc-linux-gnu, compiled by gcc (GCC) 4.8.5 20150623 (Red Hat 
> 4.8.5-36), 64-bit
>
>
>      > re: rundeck, yes - but this is only one of many things
>     connecting. it's
>      > not doing anything special.
>
>     Except sending a stream of connections to the server.
>
>     Any idea what they are doing?
>
>
> It's used to trigger ETL jobs. There are other bigger consumers of 
> connections - the issue isn't the stream of jobs, that works fine under 
> normal operation. The issue is that when jobs stay in "startup" and 
> don't enter pg_stat_activity the system spirals downwards and no new 
> connections are allowed.

Is there a correlation to these ETL jobs or the other consumers?

If so what are the consumers trying to do at that time?

> James


--
Adrian Klaver
adrian.klaver@xxxxxxxxxxx