如何在Docker容器内感知其他Docker容器的重启状态？

Hey there! Great question—this is absolutely feasible, and there are a couple of reliable ways to pull this off. Let me break down the most practical approaches for you.

方法1：Docker Events + Shell脚本（快速落地） #

This is the simplest way to get up and running quickly, using built-in Docker tools and a basic shell script.

Step 1: Grant your container access to the Docker daemon #

To let your monitoring container "see" other containers, you need to mount the Docker socket into it. This gives the container permission to interact with the host's Docker daemon:

docker run -d --name container-monitor -v /var/run/docker.sock:/var/run/docker.sock your-custom-image

Pro tip: If you want to avoid running the container as root, you can map the host's docker group ID to the container's user—just add --group-add $(getent group docker | cut -d: -f3) to the run command.

Step 2: Write the restart-monitoring script #

Create a shell script (e.g., monitor_restarts.sh) inside your container that listens for restart events and triggers your command:

#!/bin/bash
set -e

# Filter for restart events (add --filter "name=your-target-container" to watch only one container)
docker events --filter 'event=restart' | while read -r raw_event; do
    echo "[$(date)] Detected container restart: $raw_event"
    # Replace this line with your actual command
    /usr/local/bin/your-specified-command --arg1 value1 --arg2 value2
done

Make the script executable with chmod +x monitor_restarts.sh, then set it as your container's entrypoint or run it as a background process.

方法2：Docker SDK + 自定义程序（生产环境推荐） #

For more control (like filtering specific containers, handling errors, or adding logging), use the official Docker SDK for Python/Go to build a more robust listener.

Example Python implementation #

First, install the Docker SDK in your container:

pip install docker

Then write a Python script to listen for restart events:

import docker
import subprocess
from datetime import datetime

def handle_restart_event(container_name):
    print(f"[{datetime.now()}] Container {container_name} restarted! Executing command...")
    # Execute your target command (add error handling as needed)
    try:
        result = subprocess.run(
            ["/path/to/your/command", "arg1", "arg2"],
            check=True,
            capture_output=True,
            text=True
        )
        print(f"Command output: {result.stdout}")
    except subprocess.CalledProcessError as e:
        print(f"Command failed with error: {e.stderr}")

if __name__ == "__main__":
    # Connect to the Docker daemon via the mounted socket
    client = docker.from_env()
    
    # Listen for restart events (add filters={'name': 'my-app'} to target a single container)
    for event in client.events(decode=True, filters={"event": "restart"}):
        container_name = event["Actor"]["Attributes"]["name"]
        handle_restart_event(container_name)

This approach lets you add retry logic, integrate with logging systems, or extend functionality easily.

Critical Security Note #

Mounting /var/run/docker.sock gives your container root-level access to the host system. Only use this with fully trusted containers, and avoid running the container as root unless absolutely necessary.

内容的提问来源于stack exchange，提问作者Majid Mohsenifar