Prometheus and node_exporter server monitoring setup

Prometheus scrapes and stores time-series metrics. node_exporter exposes system-level data like CPU, memory and disk usage for Prometheus to collect. This guide covers installing both, configuring scrape targets, running node_exporter as a systemd service and locking down access.

Installing node_exporter

Download the latest stable release from the official GitHub releases page. As of May 2026, that's version 1.11.1. Replace amd64 with arm64 if needed.

wget https://github.com/prometheus/node_exporter/releases/download/v1.11.1/node_exporter-1.11.1.linux-amd64.tar.gz

Verify the SHA256 checksum against the value on the releases page, then extract and install:

tar -xzvf node_exporter-1.11.1.linux-amd64.tar.gz
sudo mv node_exporter-1.11.1.linux-amd64/node_exporter /usr/local/bin/

Create a dedicated system user with no home directory and no login shell:

sudo useradd --no-create-home --shell /bin/false node_exporter
sudo chown node_exporter:node_exporter /usr/local/bin/node_exporter

Test it by running /usr/local/bin/node_exporter directly, then check the output:

curl http://localhost:9100/metrics

You should see lines prefixed with node_, including metrics like node_cpu_seconds_total and node_memory_MemAvailable_bytes. By default, node_exporter exposes around 500 time series.

Running node_exporter as a systemd service

Running node_exporter from the terminal is fine for testing, but it stops when you close the session. Create a unit file at /etc/systemd/system/node_exporter.service:

[Unit]
Description=Node Exporter
Wants=network-online.target
After=network-online.target
 
[Service]
User=node_exporter
Group=node_exporter
Type=simple
Restart=on-failure
RestartSec=5
ExecStart=/usr/local/bin/node_exporter
 
[Install]
WantedBy=multi-user.target

Enable and start the service:

sudo systemctl daemon-reload
sudo systemctl enable node_exporter
sudo systemctl start node_exporter

Confirm it's running with sudo systemctl status node_exporter. The output should show active (running).

If Prometheus runs on the same host, bind node_exporter to localhost only by changing the ExecStart line:

ExecStart=/usr/local/bin/node_exporter --web.listen-address="127.0.0.1:9100"

Configuring Prometheus to scrape node_exporter

Open /etc/prometheus/prometheus.yml and add a job under scrape_configs:

scrape_configs:
  - job_name: 'node_exporter'
    scrape_interval: 15s
    static_configs:
      - targets: ['localhost:9100']
        labels:
          env: 'production'

The job_name identifies the source in queries and dashboards. targets points to the host and port where node_exporter listens. Labels like env help you filter metrics later.

Validate the config before restarting:

promtool check config /etc/prometheus/prometheus.yml

If it passes, reload Prometheus without downtime:

sudo systemctl reload prometheus

Open http://<your-prometheus-ip>:9090, go to Status > Targets and check the node_exporter job shows a green UP status. Run a quick query like node_cpu_seconds_total in the Expression Browser to confirm data is flowing.

Securing your monitoring stack

Never expose ports 9090 or 9100 to the public internet. On Ubuntu/Debian, restrict node_exporter access to your Prometheus server's IP:

sudo ufw allow from <prometheus-ip> to any port 9100

On CentOS/RHEL with firewalld:

firewall-cmd --permanent --add-port=9100/tcp

For the Prometheus web UI, put it behind a reverse proxy like Nginx with basic auth and TLS. A mesh VPN like Tailscale is another option if you need access from multiple locations without exposing ports directly.

Monitoring best practices and next steps

Use node_memory_MemAvailable_bytes instead of MemFree for memory alerts. MemAvailable accounts for buffers and cache, giving a more accurate picture of what's actually free.

Disable collectors you don't need (wifi, nfs, bcache) with the --no-collector.<name> flag to reduce noise.

For disk space alerts, the predict_linear PromQL function lets you forecast when a volume will fill up based on current trends. Setting a 7-day prediction window catches slow leaks before they become outages.

To monitor multiple servers, install node_exporter on each machine and add their IPs to the targets list in prometheus.yml. For larger environments, switch to file-based service discovery instead of hardcoding IPs.

Adding Grafana gives you visual dashboards. The Node Exporter Full dashboard (ID 1860) is a good starting point. Alertmanager routes critical alerts to Slack, email or PagerDuty.

FDC's dedicated servers and VPS plans support Prometheus and node_exporter out of the box. See FDC's dedicated server options.

Troubleshooting