I have a Catalyst 2960 acting as our DMZ switch in the network core. Lately it has begun to show very slow response time to our SNMP polling server (we use Solar Winds Orion NPM). The Solar Winds server is a VM that sits in the core along with the Catalyst. Solar Winds uses ICMP for node up/down status and then polls for SNMP information, so the response times are, I believe, ICMP echo responses. The odd thing is that I set up a constant ping from the Solar Winds server and let it run all weekend. On just plain ping responses, the Catalyst had an average response time of 2 ms back to the Solar Winds server; however, when responding to the Solar Winds polling engine, the average response time is up over 80 ms.
Of note, this is not happening with any other nodes, switches, firewalls, servers, etc. – only this one 2960 – so I don’t think there’s an issue on the Solar Winds side of the equation.
Any suggestions on what sort of logs or other data from the 2960 I could look at to see what would cause this behavior? It’s a recent development – the switch has been running in the core as our DMZ for a number of months without any issues.