How to Set Up Server Monitoring Using Prometheus and Grafana 

/ / Tutorials

Server monitoring helps you keep your infrastructure stable and responsive. Without proper insights, you often notice performance issues only after they have already affected your applications. Prometheus and Grafana work together to give you real-time visibility into your server’s CPU, memory, disk activity, and network performance. This guide explains how to install these tools, connect them, and start using ready-made dashboards and alerts. 

Understanding Prometheus and Grafana 

Before setting up anything, it helps to understand what each tool does and how they work together. 

What Is Prometheus 

Prometheus is an open-source monitoring system that collects and stores time-series data from your server. It regularly scrapes metrics such as CPU usage, memory consumption, disk I/O, and network throughput. Because the data is stored over time, you can analyze trends and identify early signs of potential issues. 

What Is Grafana 

Grafana is a visualization platform that connects to Prometheus and displays those collected metrics using customizable dashboards. Instead of reading raw numerical values, you can view graphs, charts, and gauges that update in real time. This makes it easy to monitor server behavior at a glance and react quickly to abnormalities. 

Using Prometheus and Grafana together gives you: 

  • Real-time performance monitoring 
  • Historical data analysis 
  • Customizable alerting 
  • Early problem detection 
  • Resource usage optimization 

Step 1: Install Prometheus and Node Exporter 

To get started, you first install Prometheus, which serves as your monitoring backend: 

sudo apt update 

sudo apt install prometheus 

Next, you need Node Exporter. This tool exposes your server’s hardware-level metrics so Prometheus can collect them: 

wget https://github.com/prometheus/node_exporter/releases/download/v1.6.1/node_exporter-1.6.1.linux-amd64.tar.gz tar xvfz node_exporter-1.6.1.linux-amd64.tar.gz cd node_exporter-1.6.1.linux-amd64 ./node_exporter 

After starting Node Exporter, it provides system metrics on port 9100. You can confirm everything works by opening http://your-server-ip:9100/metrics in your browser. 

Step 2: Configure Prometheus to Collect Metrics 

Now that Node Exporter is running, Prometheus needs to be told where to scrape the data. 

Open the configuration file: 

sudo nano /etc/prometheus/prometheus.yml 

Add a scrape job: 

scrape_configs: 
  - job_name: 'node' 
    static_configs: 
      - targets: ['localhost:9100'] 

Save the file, then restart Prometheus: 

sudo systemctl restart prometheus 

sudo systemctl enable prometheus 

Prometheus now collects metrics from Node Exporter at regular intervals. 

Step 3: Install and Configure Grafana 

Next, install Grafana so you can visualize your metrics: 

sudo apt-get install -y software-properties-common 

sudo add-apt-repository "deb https://packages.grafana.com/oss/deb stable main" 

wget -q -O - https://packages.grafana.com/gpg.key | sudo apt-key add - 

sudo apt-get update 

sudo apt-get install grafana 

Start Grafana and enable it at boot: 

sudo systemctl start grafana-server 

sudo systemctl enable grafana-server 

In your browser, open: 

http://your-server-ip:3000 

Log in using the default credentials: 

  • Username: admin 
  • Password: admin 

Grafana will prompt you to set a new password. 

Step 4: Connect Grafana to Prometheus 

Once inside Grafana, you connect it to Prometheus: 

  1. Go to Configuration → Data Sources 
  1. Click Add data source 
  1. Select Prometheus 
  1. Set the URL to: 
http://localhost:9090 
  1. Click Save & Test 

If configured correctly, you’ll see a green “Data source is working” message. 

Step 5: Import a Pre-Built Dashboard 

To get immediate insights, import an existing dashboard instead of building one manually. 

  1. Go to Dashboards → Import 
  1. Enter dashboard ID 1860 (Node Exporter Full) 
  1. Click Load 
  1. Choose your Prometheus data source 
  1. Click Import 

This dashboard provides comprehensive system monitoring including:  

  • CPU usage across all cores  
  • Memory utilization  
  • Disk I/O and storage usage  
  • Network traffic statistics  
  • System load averages 

Step 6: Set Up Alerting (Optional) 

By adding alerting rules, you get notifications when something goes wrong. 

You can create alerts directly inside your dashboard: 

  • Click on a panel 
  • Select Edit → Alert 
  • Define a condition, such as CPU > 80% 
  • Add notification channels like email or webhooks 

Common alert thresholds: 

  • CPU usage above 80% 
  • Memory usage above 85% 
  • Disk free space below 10% 
  • Load average higher than your number of CPU cores 

This helps you detect performance problems early. 

Understanding Your Key Metrics 

Prometheus and Node Exporter provide many useful metrics, but some are especially helpful: 

CPU Metrics 

  • cpu_usage_idle – Idle CPU time 
  • cpu_usage_system – System-level CPU usage 
  • cpu_usage_user – User-level CPU usage 

Memory Metrics 

  • node_memory_MemTotal_bytes – Total RAM 
  • node_memory_MemAvailable_bytes – Available RAM 
  • node_memory_MemFree_bytes – Free memory 

Disk Metrics 

  • node_filesystem_avail_bytes – Available disk space 
  • node_disk_read_bytes_total – Disk reads 
  • node_disk_written_bytes_total – Disk writes 

Network Metrics 

  • node_network_receive_bytes_total – Bytes received 
  • node_network_transmit_bytes_total – Bytes sent 

Knowing these values helps you understand where performance bottlenecks may occur. 

Troubleshooting Common Issues 

Prometheus not collecting metrics:  

  • Verify Node Exporter is running on port 9100  
  • Check Prometheus configuration syntax  
  • Ensure firewall allows connections between services 

Grafana can’t connect to Prometheus:  

  • Confirm Prometheus is accessible at localhost:9090  
  • Check Grafana logs: sudo journalctl -u grafana-server  
  • Verify data source configuration  

Missing data in dashboards:  

  • Allow time for metrics collection (initial data may take a few minutes)  
  • Check that your time range settings are appropriate  
  • Verify metric names match your Prometheus setup 

Advanced Configuration Options 

You can customize your setup with additional features: 

Enable extra Node Exporter collectors 

./node_exporter --collector.systemd --collector.processes 

Adjust Prometheus retention settings 

Modify prometheus.yml based on your storage capacity. 

Optimize Grafana performance 

Adjust query timeouts for large datasets. 

Add security 

Enable authentication and HTTPS for production systems. 

Additional Tips 

  • Performance: Monitor Prometheus itself to ensure it’s not consuming excessive resources  
  • Backups: Regularly backup your Grafana dashboards and Prometheus configuration 

Watch Our YouTube Video on Prometheus and Grafana 

If you prefer a visual walk-through, you can watch the connected YouTube video based on this guide. In the tutorial, you’ll see the installation steps, configuration changes, and dashboard import process visually, including the setup of Prometheus metrics and Grafana dashboards for server performance monitoring. 

Conclusion 

Prometheus and Grafana create a powerful monitoring solution that helps you track server performance in real time, identify bottlenecks early, and plan capacity upgrades. With Node Exporter supplying system metrics to Prometheus, and Grafana offering detailed dashboards and alerts, you get full visibility into your server’s health. 

You now have everything set up to monitor performance, troubleshoot issues, and stay ahead of potential problems before they affect your services. 

Scroll to Top