tdsdockerhub/docker-exporter repository overview

⁠Docker Exporter for Prometheus

This Python script (docker_exporter.py) is a simple Prometheus exporter that collects metrics from Docker containers and exposes them on an HTTP endpoint. The exported metrics cover:

• Memory: usage, limit, and working set
• CPU: percentage usage
• Network: cumulative bytes received and transmitted
• Lifecycle: restart count and OOMKilled status

Metrics are labeled with the container name and short container ID, making it easy to identify individual containers in Prometheus.

⁠How it works

1. Connects to the Docker daemon using the docker Python SDK.
2. Lists all containers (running or stopped) and for each:
   • Reloads its attributes
   • Gathers state information (restarts, OOM kills)
   • If running, fetches live stats for memory, CPU, and network
3. Updates corresponding Prometheus Gauge metrics
4. Runs an HTTP server on port 9101 for Prometheus to scrape
5. Refreshes metrics every 5 seconds

⁠Scraping with Prometheus

Add a job to your Prometheus configuration pointing at the exporter:

scrape_configs:
  - job_name: 'docker_exporter'
    static_configs:
      - targets: ['<exporter_host>:9101']

Replace <exporter_host> with the hostname or IP where the exporter is running (e.g. localhost).

Restart Prometheus after updating its config.