We're monitoring a significant number of devices via SNMP. The data is collected by a proxy group consisting of two proxies.

It is generally working but now I noticed that the polling apparently isn't distributed across all 3 SNMP poller processes. One poller consumes almost 100% of one CPU core (4 cores available) while the others apparently aren't doing anything:

I also checked the respective process states which also confirmed my assumption. The busy process is in running state while the others are sleeping.

This behaviour exists on both proxies. So that's not an individual issue.