Prometheus Node Exporter

1️⃣ What is Node Exporter?¶

Node Exporter is a Prometheus exporter that exposes OS-level system metrics from a Linux server.

It collects:¶

CPU usage
Memory usage
Disk usage & I/O
Network traffic
Load average
Filesystem stats
Hardware info

👉 It does NOT monitor applications
👉 It monitors the server itself

2️⃣ Where Node Exporter Fits¶

[ Linux Server ]
     ↓
[ Node Exporter :9100 ]
     ↓
[ Prometheus ]
     ↓
[ Grafana Dashboard ]

3️⃣ When to Use Node Exporter¶

✅ Monitor VM / bare metal
✅ Cloud servers (AWS, Azure, DO, OCI)
✅ Kubernetes nodes
❌ Not for application-level metrics

4️⃣ Default Port¶

Metrics endpoint:

http://SERVER_IP:9100/metrics

5️⃣ Installation (Linux – Recommended Way)¶

Step 1: Download Binary¶

cd /tmp
wget https://github.com/prometheus/node_exporter/releases/latest/download/node_exporter-1.7.0.linux-amd64.tar.gz

Step 2: Extract¶

tar xvf node_exporter-*.tar.gz
cd node_exporter-*/

Step 3: Move Binary¶

sudo mv node_exporter /usr/local/bin/

6️⃣ Create Systemd Service (IMPORTANT)¶

Create Service File¶

sudo nano /etc/systemd/system/node_exporter.service

Paste This¶

[Unit]
Description=Node Exporter
After=network.target

[Service]
User=nobody
ExecStart=/usr/local/bin/node_exporter

[Install]
WantedBy=multi-user.target

Enable & Start¶

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

Check Status¶

sudo systemctl status node_exporter

7️⃣ Verify Node Exporter¶

Open in browser:

http://SERVER_IP:9100/metrics

✔ If you see metrics → working correctly

8️⃣ Firewall Rules (IMPORTANT)¶

sudo ufw allow 9100

Or allow only Prometheus server IP (recommended).

9️⃣ Integrate with Prometheus¶

prometheus.yml¶

scrape_configs:
  - job_name: "node_exporter"
    static_configs:
      - targets:
          - "SERVER_IP:9100"

Reload Prometheus¶

sudo systemctl reload prometheus

🔍 10️⃣ Verify in Prometheus UI¶

Open:

http://PROMETHEUS_IP:9090

Check:

Status → Targets
node_exporter should be UP

📈 11️⃣ Grafana Dashboard (Recommended)¶

Step 1: Add Prometheus as Data Source¶

Grafana → Settings → Data Sources → Prometheus
URL: http://PROMETHEUS_IP:9090

Step 2: Import Dashboard¶

Best Dashboard ID

✔ “Node Exporter Full”

12️⃣ Key Metrics You’ll Use¶

Metric	Meaning
node_cpu_seconds_total	CPU usage
node_memory_MemAvailable_bytes	Free memory
node_filesystem_avail_bytes	Disk free
node_load1	Load avg
node_network_receive_bytes_total	Network in
node_disk_io_time_seconds_total	Disk IO

13️⃣ Alerts (Example)¶

High CPU Alert¶

- alert: HighCPUUsage
  expr: avg(rate(node_cpu_seconds_total{mode!="idle"}[5m])) > 0.8
  for: 2m
  labels:
    severity: warning
  annotations:
    summary: "High CPU usage detected"

14️⃣ Node Exporter in Docker¶

docker run -d \
  --name=node-exporter \
  -p 9100:9100 \
  --pid=host \
  -v "/:/host:ro,rslave" \
  prom/node-exporter \
  --path.rootfs=/host

15️⃣ Node Exporter in Kubernetes¶

kubectl apply -f https://raw.githubusercontent.com/prometheus/node_exporter/master/examples/node-exporter-daemonset.yaml

Runs on every node as DaemonSet.

16️⃣ Security Best Practices¶

✔ Do not expose 9100 publicly
✔ Restrict access to Prometheus IP
✔ Use firewall rules
✔ Run as non-root
✔ Use VPN / private network

17️⃣ Node Exporter vs Alternatives¶

Tool	Purpose
Node Exporter	OS metrics
cAdvisor	Container metrics
Telegraf	Metrics + logs
CloudWatch Agent	AWS only
Azure Monitor Agent	Azure only

18️⃣ Real-World Production Setup¶

Node Exporter → Prometheus → Alertmanager → Slack/Email
                         ↓
                      Grafana

19️⃣ Common Mistakes ❌¶

Mistake	Fix
Port blocked	Open 9100
Target DOWN	Check firewall
Wrong IP	Use private IP
Running as root	Use nobody

✅ Final Summary¶

✔ Lightweight
✔ Zero config
✔ Industry standard
✔ Works everywhere
✔ Perfect for infra monitoring