We're revamping our NPM/NCM environment from several stand-alone NPM regional deployments to one Main NPM server with multiple polling servers.
We just tested what we'd learn if the WAN to a regional poller failed. None of the regional sites that are monitored by it showed down.
And the poller itself did not display as a lovely and large alarm in the primary NPM node status screen; we could only see an issue was present by looking into the poller's status and seeing it hadn't been sync'd in ten minutes.
Obviously I can monitor each regional poller and send alerts they're unavailable. I already monitor the WAN routers supporting each region from the master poller in the main region's data centers, so I won't go without data.
But I'm interested in seeing what others have done to ensure they get the right alerts and displays and notification when this happens.
I sort of hoped to see a nearly overwhelming flood of down alerts coming in when a major / regional WAN goes unavailable instead of no news at all. At least then I'd know about the problem; I could build dependencies and filter out what's chaff and what's critical.
How do you handle this type of setup? Do you get alerts about the nodes that are only monitored by the unreachable NPM poller? Or no alerts at all?
Rick S.