In my mind, if a neighbor is down, then it is investigated. Our network is around 300 devices with a decent amount of OSPF peering. I know if a neighbor is broken, or if it needs to be removed after investigation. If it needs to be removed, I log into the DB and delete the neighbor in the table.
I understand link monitoring, and how I can alert on a link being up/down, but what if someone made a configuration change on a box with redundant OSPF links which broke OSPF (MTU or authentication mismatch)? If I am not alerted that OSPF is down, how do I know that my box is not operating in a redundant state? Instead of deleting them from the table, they should be changed to a
down status so they can be investigated and the appropriate action taken.
I am not going to manually create OSPF neighbor alerts for every new neighbor IP inserted into my environment, as that is just not scalable. It is much easier to make part of the process of decomissioning a retired device/neighbor IP, removing the neighbor from the DB table.
Not that it matters how Solarwinds handles this, but the way I am suggesting is exactly how Solarwinds monitors OSPF neighbors.