Skip to main content

Node Exporter

Prometheus logo

Hardware and operating system metrics from every node, ready for Prometheus.

Category: Monitoring

Node Exporter is the Prometheus project's exporter for machine-level metrics, packaged by Fireball for edge hardware (chart 2.0.3, Node Exporter 1.11.1). It reads the host's kernel and hardware counters and publishes them in Prometheus format on port 9100.

What it does​

  • Reports CPU time and frequency, load, memory, disk I/O, filesystem usage, network interface traffic, and system time.
  • Reads temperature sensors from the kernel's thermal zones, so you can see a box cooking in its cabinet before it throttles.
  • Picks up custom metrics from .prom files dropped into its textfile directory, for things like backup status or RAID health that a script can check.

When to use it​

Deploy Node Exporter when you want detailed host metrics from your edge hardware in Prometheus, for your own dashboards and alert rules.

Running on EmberNet​

  • One pod per node by default. The chart runs as a DaemonSet, so every node in the cluster gets an exporter. A single-instance mode is available as a chart option.
  • Host access, not privilege. It runs on the host network and in the host PID namespace so it can see the machine it reports on. It runs as an unprivileged user with privilege escalation off.
  • Metrics only. There is no web interface; the card shows it running, and Prometheus scrapes port 9100.
  • Lightest preset by default. The default edge-minimal preset requests 20m of CPU and 30 MiB of memory. Larger presets enable more collectors, such as hardware sensors, systemd units, and per-process metrics.

Learn more​

EmberNet packages and deploys this application; Node Exporter is developed by the Prometheus project. Product names and logos are trademarks of their respective owners.