<!--startmeta custom_edit_url: "https://github.com/netdata/netdata/edit/master/docs/npm/device-metrics/README.md" sidebar_label: "Overview" learn_status: "Published" learn_rel_path: "Network Performance Monitoring/Device Metrics" keywords: ['snmp', 'network devices', 'router', 'switch', 'firewall', 'discovery', 'interface', 'overview'] endmeta--> <!-- markdownlint-disable-file -->

Device Metrics

Netdata monitors your network devices over SNMP — routers, switches, firewalls, access points, load balancers, UPS and PDU units. It recognizes each device's model automatically and collects interface, health, and vendor-specific metrics, with no OID lists to build and no per-device dashboards to design.

Each SNMP device becomes its own node — model-recognized, with live interface, health, and vendor metrics.

What you can monitor

• Interfaces — traffic, errors, discards, and operational status, per interface.
• Device health — control-plane CPU and memory, and environmental sensors (temperature, fans, power supplies) where the device's profile exposes them.
• Vendor-specific metrics — hardware, protocol, and feature metrics from the matched vendor profile, plus BGP peers and license state where the device reports them.
• Reachability — optional ICMP latency and loss alongside every device.

Each device becomes its own node, with live charts, labels, and alerts.

Model recognition

Netdata matches each device to a profile by its sysObjectID, from a large built-in profile library. It collects standard interface and system metrics for any device that answers SNMP, and vendor-specific metrics where a profile matches; you can add custom profiles for anything special. It polls over SNMP v1, v2c, and v3.

Functions and alerts

• Live tables — snmp:interfaces (traffic, status, errors, discards), snmp:bgp-peers (per-peer state and recency), and snmp:licenses.
• Stock alerts — license expiry and BGP session health.

Where to start

• Your first device — Quick Start.
• Credentials, SNMPv3, many devices, discovery — Configuration.
• Understand or extend model recognition — SNMP Profile Format.
• Many devices or many sites — Sizing and Scaling.
• Trust the data — Validation; avoid the common mistakes in Anti-patterns.
• Something's off — Troubleshooting.