How to Install, Configure, and Monitor Node Exporter with Prometheus

Node Exporter is a Go‑written exporter that exposes *NIX host metrics such as CPU, memory, and disk without any third‑party dependencies; you only need to download, extract, and run the binary.

Installation and Configuration

Download the tarball from the Prometheus download page, extract it, and run the node_exporter binary:

wget https://github.com/prometheus/node_exporter/releases/download/v1.2.2/node_exporter-1.2.2.linux-amd64.tar.gz
# Optional fast download for China
# wget https://download.fastgit.org/prometheus/node_exporter/releases/download/v1.2.2/node_exporter-1.2.2.linux-amd64.tar.gz
 tar -xvf node_exporter-1.2.2.linux-amd64.tar.gz
 cd node_exporter-1.2.2.linux-amd64
 ./node_exporter

The log shows that Node Exporter listens on port 9100 and serves metrics at the /metrics endpoint, which can be accessed with: curl http://localhost:9100/metrics The output follows the standard Prometheus exposition format, ready to be scraped by Prometheus.

Run ./node_exporter -h to see all configurable flags. The most important flag is --collector.<name>, which enables specific metric collectors. Use --no-collector.<name> to disable defaults, or --collector.disable-defaults followed by explicit --collector.<name> entries.

For production, you can run Node Exporter in a Docker container (remember to grant access to the host’s namespaces) or manage it with systemd. Below is a typical systemd unit file:

[Unit]
Description=node exporter service
Documentation=https://prometheus.io
After=network.target

[Service]
Type=simple
User=root
Group=root
ExecStart=/usr/local/bin/node_exporter
Restart=on-failure

[Install]
WantedBy=multi-user.target

After placing the binary in /usr/local/bin, reload the daemon and start the service:

cp node_exporter /usr/local/bin/node_exporter
systemctl daemon-reload
systemctl start node_exporter
systemctl status node_exporter

Add a static scrape job for the two nodes in Prometheus configuration:

global:
  scrape_interval: 5s

scrape_configs:
  - job_name: "prometheus"
    static_configs:
      - targets: ["localhost:9090"]
  - job_name: "demo"
    scrape_interval: 15s
    scrape_timeout: 10s
    static_configs:
      - targets: ["localhost:10000", "localhost:10001", "localhost:10002"]
  - job_name: "node_exporter"
    static_configs:
      - targets: ["node1:9100", "node2:9100"]

CPU Monitoring

The metric node_cpu_seconds_total records the cumulative CPU time spent in each mode (idle, user, system, etc.) as a counter. Sample output shows values for each CPU core and mode.

To derive CPU usage percentage, calculate the proportion of time not spent in the idle mode. Using PromQL, first get the increase of idle time over a 1‑minute window, then compute the ratio against the total CPU time:

(1 - sum(increase(node_cpu_seconds_total{mode="idle"}[1m])) by (instance) / sum(increase(node_cpu_seconds_total[1m])) by (instance)) * 100

This expression yields the CPU utilization per instance, comparable to the percentages shown by the top command.

Further node‑level metrics such as memory and I/O will be covered in upcoming articles.