Distributed Monitoring

[mroter]

In a distributed monitoring environment were multiple independent Nagios XI servers are each monitoring its own site and report its check results to a central Nagios XI server(s) (running no active checks):
Who should send notifications (emails)?
(effecting the maintenance of: Contacts, contacts group, host/service escalations)
How this will compare to a Fusion or Check_MK Multisite topology?

* For availability its better that each site will be autonomic and send its own alerts
* Since the sites are very much identical it is easier to maintain it once, in the central server

Any ideas, advice?

[mguthrie]

Who should send notifications (emails)?

I would suggest the child servers, both because they're most likely running the active checks themselves, it minimizes any points of failure between the checks and the notifications. Also, if you forward checks to a central server, you want to run that central machine as lean as possible, so have notifications disabled and maybe even event handlers disabled if you can since it will free up quite a bit of CPU. However, with that said, do what works best for your environment.

How this will compare to a Fusion or Ch ... topology?
Fusion makes for a pretty simple way to setup and run a distributed monitoring environment, particularly with XI. It reduces most scalability concerns, and requires virtually no maintenance to act as an up-to-date central viewer. However, if you need to have full access to all host/service details being monitored across each site, then having a central XI would be better for you. Fusion comes with a free 60-day trial and our 2012 version has a lot of nice new features in it. I'd suggest giving it a try and see if it does what you need. If not, go for the central XI server.