The SNMP service on NSX Edge nodes stops functioning intermittently and does not recover automatically. When investigating the issue, the following symptoms are observed:
The NSX CLI command get service snmp returns the status: Service state: stopped.
The root level command /etc/init.d/vmware-snmpd status returns: Active: active (exited).
VMware NSX
This issue is resolved in VMware NSX 9.x. In this version, the service management is migrated to systemd, which includes an active watchdog to automatically recover the SNMP service.
For the NSX 4.x release line, a fix is targeted for an upcoming maintenance release.
This behavior can be reproduced by manually killing the snmpd PID (kill -9 <pid>) and observing that the service does not auto-restart.