SNMP and network devices

How to scan the network, add a switch/router/UPS and read metrics without a Windows agent.

Scan and SNMP credentials

Network devices (switch, router, UPS, network printer) often have no Windows agent. RMM → Endpoints → SNMP and network devices scans a subnet with an SNMPv2c community or SNMPv3 account.

1. Open RMM → Endpoints → SNMP and network devices (/rmm/snmp).
2. Enter IP range / subnet and community (e.g. public only in a lab) or SNMPv3 details.
3. Run the scan and wait for hosts that answer SNMP.

• Avoid default “public” in production — use a private community or SNMPv3.
• Firewalls must allow SNMP (UDP 161) from the collector / server to devices.
• No reply ≠ no device — SNMP may be off or the community wrong.

Onboard a device

From scan results pick hosts to onboard. Set a friendly name, company / location and an OID template when available.

1. Select discovered devices worth monitoring.
2. Fill a friendly name and context (company / location).
3. Save / onboard and confirm health status appears.

• Do not onboard the entire scan blindly — pick critical infrastructure.
• If the device community changes, update credentials in the panel.

Charts, interfaces and health

After onboarding, review interfaces, load, disks (when OIDs allow) and alerts. The Performance / Interfaces / Storage tabs read the latest snapshot and chart history (max 7 days; after about 24 hours, 15-minute buckets — latest walk). Full walk JSON older than about 8 days is dropped — History still shows the scan status. The /rmm/snmp list and the rules picker omit community, SNMPv3 passwords, OID lists, notes, and sysDescr — the card reads sysDescr from the latest walk. Full credentials are on the edit card. A finished collector task (agent_tasks) does not keep the community / SNMPv3 passwords or the OID list — the walk lives in snmp_data. History requests a handful of recent scans (not 100); after the API restarts the response is metadata plus one full walk (the richest successful snapshot) per page. On Dell iDRAC, the Storage (hrStorage) table is often empty — disks and RAID are on the Performance tab. For QNAP / Synology / Dell iDRAC / HPE the panel auto-creates disk/volume health and SNMP-unreachable rules on the collector (if none exist yet). QNAP volume: warning at ≤15% free, critical at ≤5% — change the threshold under Service Desk → Rules.

• Prioritize down / high-error uplinks.
• Missing charts: check credentials, timeout and OID template vs device model.
• SNMP complements Windows agents — it does not replace them.
• After adding a NAS, check Monitoring rules — you should see two auto-rules; disabling them keeps them off (no duplicates).

Collector without Worker

The network-device list is not deleted when Worker dies on the collector. Cards look offline / empty because SNMP scans are done by Worker (RMMAgent.exe); Watchdog only keeps the host online in the panel.

1. On Endpoints find the collector (e.g. VEEAM) — agent version “Worker missing”, auto-update alert.
2. On the Windows host: sc query RMMAgentWorker, file C:\Program Files\RMMAgent\worker\RMMAgent.exe, log C:\ProgramData\RMMAgent\update\watchdog-apply-*.log.
3. Restore worker\previous or reinstall the MSI. After Worker starts, scans return within a few minutes.

APC UPS (PowerNet-MIB)

The APC UPS template shows battery, load, input/output voltages and last self-test (PowerNet-MIB) on the device card. That is not the generic interface-chart view.

1. Assign the APC UPS template to the SNMP host.
2. Deploy backend and an agent that walks the OIDs, then rescan.
3. The device card should show an APC UPS section.

• An empty panel means the wrong template, no walk yet, or the scan has not returned.
• Self-test and runtime come from the UPS; Dell / QNAP / MikroTik / FortiGate / HPE cards work the same way after their own template.

FortiGate (FORTINET-FORTIGATE-MIB)

The FortiGate template shows CPU/RAM, sessions and hardware sensors. That is not the generic switch interface-chart view.

1. Assign the FortiGate template to the SNMP host.
2. Deploy backend and an agent that walks the OIDs, then rescan.
3. The device card should show a FortiGate section.

• An empty panel means the wrong template, no walk yet, or the scan has not returned.

HPE iLO (CPQHLTH-CPQIDA)

The HPE iLO template shows temperatures, fans, power supplies and physical disks (status/condition/SMART) from CPQHLTH/CPQIDA. That is not the generic switch interface-chart view.

1. Assign the HPE iLO template to the SNMP host.
2. Deploy backend and an agent that walks the OIDs, then rescan.
3. The device card should show an HPE iLO section.

• An empty panel means the wrong template, no walk yet, or the scan has not returned.

Cisco EnvMon (CISCO-ENVMON)

The Cisco template shows temperatures, fans, power supplies and CPU/RAM from CISCO-ENVMON. That is not the generic switch interface-chart view.

1. Assign the Cisco EnvMon template to the SNMP host.
2. Deploy backend and an agent that walks the OIDs, then rescan.
3. The device card should show a Cisco section.

• An empty panel means the wrong template, no walk yet, or the scan has not returned.

Related

• RMM — devices, updates, backup
• Policies, groups and maintenance windows
• Service Desk — daily operations