Module 8 Lesson 4: Debugging in Containers

When your app crashes inside a container, you can't just "Look" at it easily. You need specific tools to peer through the "Walls" of the container.

1. The "Interactive Shell" (docker exec)

This is your #1 tool. It allows you to enter a running container and run commands like ls, env, or cat to see what went wrong.

docker exec -it my-app bash
# or for small images like alpine
docker exec -it my-app sh

What to check inside:

• Config: cat .env or env (Are the variables correct?)
• Files: ls -la (Are the files in the right folder? Do they have the right permissions?)
• Connectivity: ping database (Can this container see the DB?)

---

2. Advanced Logging

If your app won't even start, you can't exec into it. You must use logs.

• docker logs <id>: Show the history.
• docker logs --tail 50 <id>: Just the last 50 lines.
• docker logs -t <id>: Show Timestamps (Critical for syncing app logs with database errors).

---

3. Remote Debugging (Connecting your IDE)

Modern IDEs (VS Code, GoLand, PyCharm) allow you to connect their Visual Debugger to a process running inside a container.

1. Expose a Debug Port: Launch your container with an extra port (e.g., -p 9229:9229 for Node.js).
2. Start in Debug Mode: Change your app's start command (e.g., node --inspect=0.0.0.0 app.js).
3. Attach IDE: In VS Code, use a launch.json configuration to "Attach" to that port.
4. Profit: You can now set Breakpoints in your local code, and the container will "Pause" when it hits them!

---

4. Inspecting Metadata

Sometimes the problem isn't the app; it's the Docker configuration.

docker inspect <id>

Look for:

• Mounts: Is the volume actually mounted to the right place?
• NetworkSettings: What is the IP address?
• State: Why did it exit? (Check the Error and ExitCode).

---

Exercise: The Mystery Crash

Scenario: Your Python app keeps exiting with a ModuleNotFoundError.

1. List all containers (including stopped ones). Found it?
2. Run docker logs. What is the exact name of the missing module?
3. "Jump inside" a temporary instance of that image: docker run -it my-image sh.
4. Run pip list. Is the library there?
5. Check your Dockerfile. Did you forget to COPY requirements.txt before running pip install?
6. How would you use docker inspect to verify that your requirements.txt was actually copied into the /app folder?

---

Summary

Debugging in Docker is a two-step process: First, look at the Metadata (Logs and Inspect). Second, look at the Internal State (Exec). Mastering these tools turns "Invisible bugs" into solvable problems.

Next Lesson: Final polishing: Development vs. Production Dockerfiles.