There’s no doubt that e-mail is a vital component in business today, and as such we pay particular attention to our e-mail infrastructure.
We use dedicated systems as Mail Transfer Agents, of course, and run Exim on those, with as little ballast as possible. What we do though, is to give each MTA its own DNS caching server, and generally try to ensure all services the MTA requires are close to it, network wise.
- We built a light-panel in Nagios Blinkenlights, which greatly impressed the management. (You may even be able to get that extra bit of budget you need for …)
- Use whatmon in your Firefox browser or in Thunderbird to keep abrest of things going wrong. If you don’t typically use these Mozilla programs, there are workarounds like using GeekTool on a Mac.
- Even if your setup is convoluted, there is nothing I can think of that, with a bit of work, Nagios can’t monitor. I even extended Nagios monitoring via XML-RPC to reach the back of beyond.
- Get Nagios to add issues to your trouble-ticket system: we did that with Mantis.
- We monitor ClamAV, to ensure its database is up to date.
- The ultimate Nagios monitor is Naguino: an Arduino-based LCD monitor, of course. But I’m biased. :-)
- Do you use DNS black-lists? If so, periodically test your MTA against them.
- There is nothing that cannot go wrong; keep your monitoring stuff up to date, and whenever some problem crops up, add a check for it to Nagios. (I learnt the hard way, which is why I now also check for expired CRLs.)
Giving your monitoring environment a bit of thought (and quite a bit of work) quickly pays off. It is a relief to be able to look at your monitoring panel (or Web page) and see something like this: No red. No Yellow. All green. The best there is! ;-)