Why import Null or Loopback?

I wouldn't bother monitoring the Loopback or Null interfaces. I certainly don't. Just like I don't monitoring user interfaces using of course has complained about a 'network issue'. Uplinks are worth monitoring though.

Right, that is my feeling I just want to make sure I'm not overlooking something that some advanced sniffer tech would say I was insane to have turned off loopback monitoring.

I have decided not to monitor either but there was a time when due to interfaces changing etc. the one interface you would always be able to reach on a device would be the loopback since it doesn't change and can be a subnet selected just for management.

I think there is a difference here being monitoring and polling.

We POLL the device via the loopback address, however we do not MONITOR the loopback interface for statistics...

Why would I add the loopback to poll and burn an element versus the IP address of the system itself? That would be several thousand elements in my environment.

It wouldnt burn an element as you'd use the loopback as the system ip address, instead of the address on one of the interfaces.

There's a good blog posting about this by Josh Stephens here which explains it far better than I could

Thanks stuartwhyte!

Also, as it is written in the blog post - loopbacks interfaces make sense mainly for routers or devices with more physical interfaces with different IP on each. If you have just a bunch of end-stations with single physical interface and IP address then I would not use their loopback to be monitored via NPM - you should be able to easily filter it via new interface filter page in 10.5.