Simple Docker Agent

A lightweight Python-based agent that provides real-time monitoring of Docker containers and host system resources.

Features

• REST API: Exposes a /status endpoint for easy integration.
• Container Insights: Reports CPU, memory (raw and humanized), uptime, ports, labels, and more for each container.
• Host Metrics: Shows CPU, memory, and disk usage for the host (raw and humanized).
• Fast & Lightweight: Designed for quick responses, even with many containers.
• Easy Deployment: Runs as a Docker container with minimal configuration.

Screenshot

---

Quick Start

Run from Docker Hub (method 1)

You can run the latest public version of Simple Docker Agent directly from Docker Hub, with no need to build the image yourself.

docker pull solutionforest/simple-docker-agent:latest

docker run -d \
  --name simple-docker-agent \
  -p 8080:8080 \
  -v /var/run/docker.sock:/var/run/docker.sock \
  --privileged \
  solutionforest/simple-docker-agent:latest

• The agent will now be accessible at http://localhost:8080/status.
• Make sure to mount the Docker socket (-v /var/run/docker.sock:/var/run/docker.sock) to allow the agent to access container information.
• The --privileged flag may be required for collecting some host metrics.

You can also specify a version instead of latest, for example:

docker pull solutionforest/simple-docker-agent:1.0.0
docker run -d \
  --name simple-docker-agent \
  -p 8080:8080 \
  -v /var/run/docker.sock:/var/run/docker.sock \
  --privileged \
  solutionforest/simple-docker-agent:1.0.0

See solutionforest/simple-docker-agent on Docker Hub for available tags and more info.

1. Build the Agent Docker Image (method 2)

git clone https://github.com/solutionforest/simple-docker-agent
cd simple-docker-agent
docker build -t simple-docker-agent .

2. Run the Agent

docker run -d \
  --name docker-agent \
  -p 8080:8080 \
  -v /var/run/docker.sock:/var/run/docker.sock \
  --privileged \
  simple-docker-agent

3. Query the Status API

curl http://localhost:8080/status | jq

---

Example Response

{
  "host": {
    "cpu_percent": 4.5,
    "memory": {
      "total": 16777216000,
      "used": 2548039680,
      "percent": 15.2,
      "total_human": "16.8 GB",
      "used_human": "2.5 GB"
    },
    "disk": {
      "total": 128849018880,
      "used": 64424509440,
      "percent": 50.0,
      "total_human": "128.8 GB",
      "used_human": "64.4 GB"
    }
  },
  "docker_services": [
    {
      "id": "abcdef123456...",
      "name": "my-service",
      "status": "running",
      "image": ["nginx:latest"],
      "stats": {
        "cpu_percent": 0.21,
        "cpu_human": "0.21%",
        "mem_usage_bytes": 25600000,
        "mem_usage_human": "25.6 MB",
        "mem_limit_bytes": 2147483648,
        "mem_limit_human": "2.1 GB"
      },
      "created_at": "2024-05-01T12:00:00Z",
      "started_at": "2024-05-01T12:01:00Z",
      "uptime_seconds": 3600,
      "uptime_human": "1 hour",
      "restart_count": 0,
      "ports": {
        "80/tcp": [
          {
            "HostIp": "0.0.0.0",
            "HostPort": "8081"
          }
        ]
      },
      "labels": {
        "com.docker.compose.project": "myproject"
      }
    }
    // ...more containers
  ]
}

---

Web Dashboard

A modern, interactive dashboard UI is included! Just open:

http://localhost:8080/dashboard.html

in your browser after starting the agent. No extra setup is required.

Dashboard Features

• System Overview: Uptime, CPU idle %, core count, memory, swap, and disk usage cards.
• Charts: Real-time graphs for system load, CPU modes, memory, disk IO, network, and swap.
• Container Selector: Switch between running containers to view their stats.
• Per-Container Stats: Status, image, uptime, restart count, ports, and resource usage (CPU %, memory MB) with historical charts.
• Responsive Design: Works on desktop and mobile.

---

API

• /status: Returns a JSON object with host and container stats. See example above for structure.
• /dashboard.html: Serves the dashboard UI.

---

Running Without Docker (Advanced)

You can run the agent directly with Python 3.8+ (Linux recommended):

pip install flask psutil docker humanize
python agent.py

Then visit http://localhost:8080/dashboard.html in your browser.

---

Configuration

• Port: Default is 8080 (see agent.py).
• Docker socket: Mount /var/run/docker.sock for API access.
• Privileges: The container may require --privileged for resource stats.

---

Security

Do not expose the agent’s API to the public internet without authentication or firewall protection.
For production deployments, consider running behind a reverse proxy and securing the endpoint.

---

License

MIT License

---

Credits

• psutil
• docker-py
• humanize
• Flask

---

Contributing

Pull requests and feature suggestions are welcome!