Prometheus NUT Exporter

A Prometheus/OpenMetrics exporter for uninterruptable power supplies (UPSes) using Network UPS Tools (NUT).

Usage

NUT

Set up NUT in server mode and make sure the TCP port (3493 by default) is accessible (without authentication).

If you want to test that it's working, run telnet <nut-server> 3493 and then VER, LIST UPS and LIST VAR <ups>.

Docker

Example docker-compose.yml:

services:
  nut-exporter:
    image: ghcr.io/seb26/prometheus-nut-exporter:v1.2.2
    environment:
      - TZ=Europe/Oslo
      - HTTP_PATH=/metrics
      # Defaults
      #- RUST_LOG=info
      #- HTTP_PORT=9995
      #- HTTP_PATH=/nut
      #- LOG_REQUESTS_CONSOLE=false
      #- PRINT_METRICS_AND_EXIT=false
    ports:
      - "9995:9995/tcp"

Prometheus

Example prometheus.yml:

global:
    scrape_interval: 15s
    scrape_timeout: 10s

scrape_configs:
  - job_name: "nut"
    static_configs:
      # Insert NUT server address here
      - targets: ["nut-server:3493"]
    relabel_configs:
      - source_labels: [__address__]
        target_label: __param_target
      - source_labels: [__param_target]
        target_label: instance
      - target_label: __address__
        # Insert NUT exporter address here
        replacement: nut-exporter:9995

In the above example, nut-exporter:9995 is the address and port of the NUT exporter while nut-server:3493 is the address and port of the NUT server to query through the exporter.

Kubernetes Resource Usage

Example container resources requests and limits. This was done by scraping one NUT server with two UPSes. Resource usage was observed across 7 days period. This can be lowered even more but should be sufficient as a starting point.

resources:
  limits:
    cpu: "10m"
    memory: "16Mi"
  requests:
    cpu: "1m"
    memory: "8Mi"

Grafana

Example dashboard

Configuration

Docker Image Versions

Images are published to GitHub Container Registry, at `ghcr.io/seb26/prometheus-nut-exporter.

Use the exact version number for releases (i.e. v1.2.2) and latest for bleeding/dev/unstable releases.

Environment Variables

• RUST_LOG (defaults to info): The log level used by the console/STDOUT. Set to debug so show HTTP requests and trace to show extensive debugging info.
• HTTP_ADDRESS (defaults to ::): The HTTP server will listen on this IP. Set to 127.0.0.1 or ::1 to only allow local access.
• HTTP_PORT (defaults to 9995): The HTTP server port.
• HTTP_PATH (defaults to nut): The HTTP server metrics path. You may want to set it to /metrics on new setups to avoid extra Prometheus configuration (not changed here due to compatibility).
• PRINT_METRICS_AND_EXIT (defaults to false): Print a Markdown-formatted table consisting of all metrics and then immediately exit. Used mainly for generating documentation.
• UPS_POWER_FROM_LOAD_PERCENTAGE (defaults to false): when set to true and a value for ups.power is not available from a UPS, an approximate value will be calculated for the metric nut_power_watts (Apparent Power). Useful for UPS units that do not report the load in Watts, like some CyberPower units. The calculation will be: ( ups.load / 100.00 ) * ups.realpower.nominal.

Metrics

See metrics.

License

GNU General Public License version 3 (GPLv3), authored by HON95, originally available at https://github.com/HON95/prometheus-nut-exporter.

This is a fork by seb26, made because the developer isn't active anymore.

It is forked from v1.2.1 and the license is the same.

The first release number from this fork is v1.2.2, and includes work by sirux88 and seb26. Please see the repository history for more detailed ownership information.

I like the program and honour the original author's work (HON95), but I just need it to be up to date and have patches.

Isn't that what open source is all about?

As with most open source software from GitHub, please use at your own risk.