Installer node_exporter avec Homebrew, dnf, MacPorts, zypper

Liens du projet

• Page d'accueil
• Dépôt
• Docs amont
• Page du gestionnaire de paquets

Commandes et alias

• node_exporter
• node_exporter_brew_services

Historique du projet

The GitHub repository was created on 2013-04-18 under the Prometheus organization. The README describes it as a Go exporter for hardware and OS metrics exposed by *NIX kernels, with pluggable collectors and a separate recommendation to use `windows_exporter` for Windows machines.

The project evolved by growing a large collector surface while keeping some expensive or high-cardinality collectors disabled by default. The README documents default collectors such as cpu, diskstats, filesystem, loadavg, meminfo, netdev, textfile, uname, vmstat, and many others, plus opt-in collectors such as systemd, perf, processes, tcpstat, wifi, and zoneinfo.

Historique d'adoption

By 2026-07-01, GitHub metadata reported 13562 stars and 2672 forks for `prometheus/node_exporter`. Homebrew's formula API reported stable version 1.11.1 and 5544 installs over 365 days. That Homebrew number is only a small slice of usage because production deployments often use distro packages, containers, configuration management, or upstream release tarballs.

The Prometheus documentation treats Node Exporter as the standard path for Linux host metrics and walks through downloading a release, running the binary, scraping `localhost:9100`, and checking metrics with the `node_` prefix. Grafana and Prometheus ecosystem docs commonly use it as the base host telemetry example.

Modes d'utilisation

The practical pattern is boring and important: run one `node_exporter` per host, expose port 9100 to Prometheus, and tune collectors only when needed. In containers, the README warns that the exporter is designed to monitor the host system, so Docker deployments bind-mount the host root filesystem, use host networking and PID namespace, and set `--path.rootfs=/host`.

Package nerds care about collector flags. Default collectors cover the normal node dashboard path; opt-in collectors need testing because scrape duration, timeout, and sample cardinality can jump. The textfile collector is the escape hatch for local scripts and cron jobs: write Prometheus-format files to a configured directory and let node_exporter expose them with the rest of the host metrics.

Pourquoi les passionnés de paquets s'y intéressent

This is not related to Node.js despite the name. In Prometheus vocabulary, a node is a machine. That naming collision is worth encoding because package search and enrichment can otherwise confuse JavaScript tooling with host observability tooling.

Chronologie

• 2013-04-18: GitHub repository `prometheus/node_exporter` was created.
• 2024-07-14: v1.8.2 release metadata appeared in GitHub releases.
• 2025-10-24: v1.10.0 was released.
• 2026-04-07: v1.11.1 was released and Homebrew reported it as the stable formula version on 2026-07-01.

Related projects

• Prometheus
• Grafana
• windows_exporter
• prometheus-community/node-exporter-textfile-collector-scripts

Sources

• https://api.github.com/repos/prometheus/node_exporter
• https://formulae.brew.sh/api/formula/node_exporter.json
• https://github.com/prometheus/node_exporter
• https://github.com/prometheus/node_exporter/releases
• https://prometheus.io/docs/guides/node-exporter/
• https://raw.githubusercontent.com/prometheus/node_exporter/master/README.md