Skip to content

Device added but SNMP won't poll

You add an OLT or switch to NetSense, availability (ping) turns green, but no SNMP data arrives: no traffic, no ONUs, no inventory details. The device typically sits in Warning with “SNMP timeout” style alarms.

This is the single most common onboarding issue we see, and in almost every case the device itself is fine.

An access list (ACL) on the device permits SNMP only from specific IPs, and the NetSense agent’s own IP is not one of them.

The subtle part: NetSense polls from the agent VM’s local IP address, not from a NetSense cloud address. Firewall rules are often opened for our public ranges during setup, but the device’s own SNMP ACL never learns the agent’s private address, so the device silently drops the agent’s SNMP requests while still answering its ping.

  1. Add the agent VM’s IP address to the SNMP access list of every monitored device. If you prefer a range, a small subnet (e.g. a /30) covering the agent is enough.
  2. On VSOL and Syrotech OLTs the relevant page is System Configuration → Login Management in the web UI.
  3. To isolate the problem, you can temporarily disable the device’s SNMP ACL, confirm data starts flowing, then re-add a scoped permit for the agent’s IP.

While you’re there, double-check the basics that produce the identical symptom:

  • Community string typos: the community configured in NetSense must match the device exactly (we have genuinely seen pubic vs public).
  • SNMP version: NetSense polls SNMP v2c. A device configured for v3 only will time out. v2c plus an ACL restricted to the agent IP is the recommended setup. (NetSense does not use SNMP traps, so no trap configuration is needed.)
  • OLTs on public IPs: if the device is reached through a public IP, confirm UDP port 161 is forwarded from the public address to the device.

Within a few minutes of fixing the ACL, the device page should start filling in: model and firmware in the inventory, traffic on ports, and (for OLTs) the ONU list. If ping works and SNMP still times out after the ACL change, re-check which source IP the device actually sees and confirm the community, then contact support: we can test from the agent side in minutes.