Some switches/routers monitored via SNMP v3 are stuck as unavailable if they drop off for any reason, such as temporary network connection problems. The equipment does not need to be rebooted for the issue to surface, an external reason is enough.

The problematic devices tend to be somewhat large and/or busy. Most often this issue hits two different HPE 5900AF-48XG-4QSFP+ switches, which are in our datacenters.

This is a follow-up from ZBX-9123.

ZBX-8385 may be related, but neither of its initial cases a) or b) apply - engineID's are unique, the devices are not rebooted and their boot counts increment.