DevOps Daily

/home

/articles

/topics

Home > Articles > Analyzing Physical Server Usage with Grafana and Prometheus

Analyzing Physical Server Usage with Grafana and Prometheus

  ·   4 min read

#devops #observability #grafana #prometheus

Analyzing Physical Server Usage with Grafana and Prometheus

In modern DevOps environments, understanding server performance is crucial for maintaining reliability and optimizing resource usage. When it comes to monitoring physical servers, two powerful open-source tools—Prometheus and Grafana—form an excellent combination. Prometheus acts as the data collection and alerting engine, while Grafana is used to visualize the data, making it easier for teams to analyze and respond to changes in server performance.

This article will guide you through the setup and integration of Prometheus and Grafana for monitoring physical server metrics such as CPU usage, memory, disk I/O, and network traffic.

Why Use Prometheus and Grafana?

  • Prometheus: Prometheus is a robust monitoring system that scrapes metrics from configured targets (such as servers, containers, or applications) at specified intervals. It stores these metrics and provides a query language, PromQL, to retrieve and analyze the data. It is well-suited for time-series data, making it ideal for server monitoring.
  • Grafana: Grafana is a powerful open-source platform for creating and sharing dashboards. It integrates seamlessly with Prometheus, allowing users to visualize metrics, set up alerts, and build custom dashboards for a clear overview of server health.

Setting Up Prometheus for Physical Server Monitoring

Before starting, make sure you have access to your physical server and permissions to install software. Here’s a step-by-step guide:

1. Install Prometheus on Your Physical Server

  1. Download Prometheus from the official website: 
    wget https://github.com/prometheus/prometheus/releases/download/v2.46.0/prometheus-2.46.0.linux-amd64.tar.gz
  2. Extract the downloaded file: 
    tar -xvf prometheus-2.46.0.linux-amd64.tar.gz
cd prometheus-2.46.0.linux-amd64
  3. Configure Prometheus to start as a service or run it manually: 
    ./prometheus --config.file=prometheus.yml

2. Configure Node Exporter to Collect Server Metrics

Prometheus needs a way to collect data from your physical servers. For this, you’ll use the Node Exporter, a Prometheus exporter for hardware and OS metrics.

  1. Download and install Node Exporter: 
    wget https://github.com/prometheus/node_exporter/releases/download/v1.6.0/node_exporter-1.6.0.linux-amd64.tar.gz
tar -xvf node_exporter-1.6.0.linux-amd64.tar.gz
cd node_exporter-1.6.0.linux-amd64
  2. Start Node Exporter: 
    ./node_exporter
  3. Modify the prometheus.yml file to include Node Exporter as a scrape target: 
    scrape_configs:
  - job_name: 'node_exporter'
    static_configs:
      - targets: ['localhost:9100']

3. Verify Prometheus Configuration

Access the Prometheus web UI by visiting http://<your-server-ip>:9090. Use PromQL queries to check if Node Exporter metrics are being scraped:

node_cpu_seconds_total
node_memory_MemTotal_bytes

Visualizing Metrics with Grafana

Once Prometheus is collecting data, the next step is to set up Grafana to visualize these metrics.

1. Install Grafana

  1. Download and install Grafana: 
    sudo apt-get install -y adduser libfontconfig1
wget https://dl.grafana.com/oss/release/grafana-11.0.1.linux-amd64.tar.gz
tar -zxvf grafana-11.0.1.linux-amd64.tar.gz
cd grafana-11.0.1/bin
./grafana-server
  2. Access Grafana via http://<your-server-ip>:3000 and log in using the default credentials (admin/admin).

2. Add Prometheus as a Data Source in Grafana

  1. Go to Configuration > Data Sources and click Add data source.
  2. Select Prometheus and enter the URL for your Prometheus server (e.g., http://localhost:9090).
  3. Click Save & Test to verify the connection.

3. Create Dashboards to Visualize Server Metrics

With Prometheus configured as a data source, you can now create custom dashboards:

  1. Import Pre-Built Dashboards: Grafana has a library of pre-built dashboards. Search for popular Node Exporter dashboards by visiting Grafana Dashboards.
  2. Create Custom Dashboards:
    • Click on New Dashboard and add Panels.
    • Use queries like: 
      node_cpu_seconds_total{mode="idle"}
node_memory_Active_bytes
node_disk_io_time_seconds_total
node_network_receive_bytes_total
    • Customize the panels by adjusting time intervals, axes, and display options.

Example Use Cases

  1. CPU Usage Monitoring: Use PromQL to get CPU utilization over time. Visualize the average CPU load and identify trends or unusual spikes.
  2. Memory Usage: Track available and used memory. Analyze memory allocation patterns to decide when to scale up server resources.
  3. Disk I/O: Monitor disk read/write speeds to detect potential bottlenecks, especially in database-heavy environments.
  4. Network Traffic: Measure incoming and outgoing network traffic to ensure that network bandwidth is not a limiting factor.

Setting Up Alerts

One of the benefits of using Prometheus with Grafana is the ability to set up alerts. Prometheus can be configured to trigger alerts when certain conditions are met (e.g., CPU usage exceeds 90% for more than 5 minutes).

  1. Define Alert Rules in prometheus.yml: 
    alerting:
  alertmanagers:
    - static_configs:
        - targets: ['localhost:9093']
alert_rules:
  - alert: HighCPUUsage
    expr: node_cpu_seconds_total{mode="system"} > 0.9
    for: 5m
    labels:
      severity: critical
    annotations:
      summary: "High CPU Usage Detected"
      description: "CPU usage is above 90% for the last 5 minutes."
  2. Set Up Alerting in Grafana: Go to Alerting > Contact Points and configure how you want to receive alerts (e.g., email, Slack, PagerDuty).

Conclusion

Prometheus and Grafana provide a robust and flexible solution for monitoring physical servers. With Node Exporter, Prometheus scrapes metrics, while Grafana visualizes this data for easy analysis. The combination of these tools helps teams stay informed, anticipate issues, and make data-driven decisions to optimize server performance. By setting up alerts, you can proactively address issues before they become critical, ensuring reliability and performance across your infrastructure.

References

This guide should give you a solid foundation for setting up monitoring for physical servers using these tools. Happy monitoring!

←
Challenges of Observability in DevOps: An Expert Perspective
Using RabbitMQ for Messaging with a Homemade ESP32 Sensor
→
back to top