Monitor BGP Downtime With NPM

I created a node for the virtual IP shared between our two BGP routers and use that along with a couple of legacy graphs on the classic dashboard to keep an "always on" view into it but you can also apply alerting rules to that specific node for the interface. Out of box NPM should have already set up most of the basics.

You may get more granular information from the Voip & Network Quality WAN monitors (don't currently use that module so can't explicitly say)

That sounds like a good idea! It would allow us to track BGP neighborship if they are reachable to one another at least. There are some cases where the BGP gets stuck in 'active' state even though the two neighbor IPs can see one another and one BGP session on either end needs to be cleared, so I don't know that it would matter if the neighbor IPs can reach one another in that monitor.

The other status to check that I think would be more beneficial would be to watch for a sudden downward spike in traffic. I don't think that would be easily accomplished. Looking back at the traffic graphs when we have lost neighborship or a BGP connection the 'Min/Max/Avg bps in/out' graph still shows traffic but just very little. Perhaps we would need to watch the 'Percent utilization' drop down to something like 0% and alert on that as we should always be seeing a higher value - but how can we do that for just this single interface?