I have brought up something like this before But I wanted to mention it again but from a different angle.
One of the biggest use cases I have for NPM is to monitor the "health" of large chassis switches and routers (3800/3900--4500-6500--nexus 7k/5k). Any place I have one of these there will be hundreds of people depending on them for connectiivty. I do this now primarily through syslog which has served us well but there is no static stats or history to look at for trending or any kind of planning. I think something like this could be critical.
if I wanted to go that route I could save thousands and just download MRTG.
What I would like to have is the option to click on a chassis deivce (or a nonchassis devices if possible). Then you would get a pic showing dots/info connected to or in front of each important peice. So for example I click Distro-1 from manage nodes then I can see all te dots on things like the power supplies , important interfaces like uplinks , temps, cpu mem etc etc. And they should go from green for OK, to yellow for wanring to red for critical.
Obiously alerts here would be the best thing but seeing something like this gives you a lot more options that take less time to see what's going on on a chassis. This has a lot of value for normal times andas far as sanity checking your alerts and planning. It would become invaluable during a time of crysis.
