postfix monitoring

[deln...@gmail.com]

Considering the best practices, and the fact that Postfix is running by out of the box on almost all popular distros - is running Postfix exporter along with Node exporter regular setup now days?  Sounds like an overkill since most of these instances are not actual mail servers.

Alternative approach would be running the bash script that would expose the status through Node exporter text collector?

[Brian Candler]

It depends on whether you care about the status of postfix itself.  For systems which are just running postfix as a replacement for /usr/bin/sendmail, then it's only being used for outbound deliveries - e.g. cron and error reports.  In this instance, I suppose it could be interesting to measure the size of the queue, and count rejected mail - as you say, a textfile exporter would be straightforward.  Alternatively, using the systemd collector of node_exporter may be good enough to check that the postfix service is running.

Personally I don't run MTAs on servers in general: I install ssmtp and point it to a local mail relay.  Then there's nothing to monitor, apart from the central mail relay itself.  This also scales better with containers - if you have 20 containers on one machine you don't want 20 MTAs running.