Nagios Support Forum

Hello, we've got a number of data centres and until now we've gotten by with having a nagios server in each DC (more or less) but we're expanding and getting more instances where we have odd servers here and there outside of our DCs. For example, we've got a server in australia and often enough the nagio server that monitors it alerts. There's nothing wrong with the server in australia, it's just the route from here to there craps out.
Sure we could have a nagios server in australia, but if we monitor that, we get the same issue and if we don't monitor that then we run the risk of missing alerts because the monitoring server has died and we didn't notice.
In any event, we're getting more stuff like this, so that approach will lead to us having dozens of independent nagios servers (we've already got five and that's too many) so I'm looking to consolidate into a more centrally managed setup

Specifically I'm looking for something wherby nagios will do all its usual checks, then if it detects a problem, it will check that service again from another node, if the second node concurs, then it sends an alert.

I was looking into DNX, then I noticed it hasn't been updated in 3 years. I've also been looking at mod_gearman and merlin but frankly I'm not sure if they can actually do the thing I need.

Can anyone reccommend either of these or any sort of alternate solution?

vmos,

Have you considered a passive solution to your servers that are outside the DC? For example, if you have passive agent sending data back from the "Australia" node, if the Australia node can't send it's metrics back nothing would happen. If it does get through, Nagios will evaluate the check results sent from the Australia node. Passive checks on Nagios Also have a "freshness" setting in which to take action if it has not heard from the node for an extended period of time.

This might do what you're looking for without all of the redundant checking.


-Yancy

thats a tricky one, we've been discussing it internally and will need a bit of time to come up with a solid idea for you. If no one beats us to it, I would look for something tomorrow with a good start from us!

Thanks Yancy, I wasn't aware of the freshness setting, but wouldn't that setup require duplicate instances of nagios with duplicate configs?

vmos,

I'm not sure why you would require duplicate configs and duplicate instances. In Australia, you would have zero instances of Nagios, but have that server in Australia being monitored submit results to your DC.

Can you expand on why you think you need duplicate instances and configs.. Maybe i'm missing something.


Thanks,

-Yancy