Skip to content

What Is a Collector?

A collector is a small, lightweight agent you run inside your network. It’s how GridNMS actually sees your equipment. Every GridNMS deployment — cloud or self-hosted — uses at least one.

  • Checks reachability — pings your devices to track up/down status.
  • Polls for metrics — gathers interface traffic, CPU, memory, disk, and more from your devices using the credentials you provide.
  • Receives logs and traps — listens for syslog and SNMP traps sent by your devices.
  • Runs discovery scans — finds new devices on the ranges you assign.

It then sends all of this back to GridNMS over a single secure outbound connection.

You never have to open inbound firewall ports to GridNMS. The collector always makes the connection out to your GridNMS instance, which is simpler and safer — especially across sites and through NAT.

Your site GridNMS
┌─────────────────────┐ ┌──────────────┐
│ switches / routers │ │ │
│ firewalls / servers │──► collector ─┼──► web app │
│ syslog + SNMP traps │ outbound, │ dashboards │
└─────────────────────┘ encrypted └──────────────┘

Run one collector per site or network segment. A single collector can watch many devices across the networks you assign to it, so a small office needs just one. Add more for:

  • Additional physical sites — one collector per location.
  • Isolated network segments — when a single collector can’t route to every subnet.
  • Scale and resilience — spread load across several collectors.
Your setup Collectors
One office 1
Headquarters + 3 branches 4 (one each)
Self-hosted, single site The built-in collector that ships with the stack
Self-hosted, multi-site The built-in one + one per remote site

A collector only talks to the networks you assign to it (plus your GridNMS instance). This assignment is the boundary that controls what it can scan and monitor — see Networks & Sites.

Collectors update themselves automatically over their existing connection to GridNMS. There’s nothing to patch by hand.