This more looks like temporary network swinging. If you don't have, I would suggest you to build the dependencies (manually or you can try automatic as well) between the servers in relation to the polling engine (e.g. You know, that server A, B and C went down, but the servers B, C have the server A on the way to the Polling engine, so B and C will be the dependency childs). The dependency should then cause, that if the dependency parent goes down and the children will be unpingable as well, those will go to the Unreachable state. This will hopefuly help you to figure out which network element was the swinging one (ideally when dependencies are built correctly, this will be the only one who went down).
↧