I would to be able to rapidly detect and record connectivity problems between machines in different subnets. Kind of similar to the way BFD can help you achieve this over transit links.
There's the old classic, continuous "ping" test. But can we do better than that? Ping sometimes gives deceptive results during asymmetric routing, and ICMP is sometimes deprioritized on network gear (and not really representative of User Experience anyway: 1 second polling intervals on a protocol they don't even use..)
Does anyone have any recommendations on how to monitor host machines with that level of granularity?
No comments:
Post a Comment