From: Tom Lane <tgl@sss.pgh.pa.us>
To: Mark Steben <mark.steben@drivedominion.com>
cc: pgsql-admin <pgsql-admin@postgresql.org>
Subject: Re: sysout error log being truncated
In-reply-to: 
 <CADyzmyyftHOPk_Tc0MpdYqGKCZ2KeVaqv3rRci5NRYH-+7CTWA@mail.gmail.com>
References: 
 <CADyzmyyftHOPk_Tc0MpdYqGKCZ2KeVaqv3rRci5NRYH-+7CTWA@mail.gmail.com>
Comments: In-reply-to Mark Steben <mark.steben@drivedominion.com>
	message dated "Fri, 08 Mar 2019 17:22:16 -0500"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <22233.1552086577.1@sss.pgh.pa.us>
Date: Fri, 08 Mar 2019 18:09:37 -0500
Message-ID: <22234.1552086577@sss.pgh.pa.us>
Precedence: bulk

Mark Steben <mark.steben@drivedominion.com> writes:
> One of our transactions is a queue processor that generates a ton of insert
> statements per run of this transaction - as many as 1000 per run.
> What has happened in new datacenter is that these transactions - while
> running correctly and completely - do not write all the insert statements
> to our sysout log - only some.

Most, if not all, implementations of syslog will drop messages under
sufficient load; I think they consider that a feature.  It's not Postgres'
fault and there's nothing we can do about it.  Use a different logging
mechanism if you can't afford to lose any messages.

			regards, tom lane