We would like to monitor mail queues on our machines.
We use postfix.
On the vast majority of systems the queue should be empty almost all the time. (Only cron or admin emails would ever be in them).
The exceptions to this are:
bastion has all our outgoing email on it. So the queue here often is several thousand deferred for various reasons (remote hosts down, mailboxes full, etc).
smtp-mm0* machines are our mailmain frontends. They could have emails deffered on them.
So, the several steps here might be:
Gather info on machines and how many emails they "normally" have in their queue.
See if there's a way to note when this activity becomes abnormal. Possibly a large number of defers in a short time, or the overall queue reaching something very high.
Test.
implement checks.
First I found this plugin, that could me help:
http://exchange.nagios.org/directory/Plugins/Email-and-Groupware/Postfix/check_postfix_queue/details
I will do a test.
attachment Add_check_mail_queue.patch
Patch added, please take a look.
Patch looks fine from a quick glance.
I would not be opposed to adding a collectd check for this too, so we can take stats and maybe tune our monitoring to be a bit system-specific, or service-specific.
Check working well.
Login to comment on this ticket.