How I turned Docker incidents into game levels (with postmortems)

A tutorial typically shows you how to start a container. However, it rarely teaches what to do when a container goes down at 3am and users are waiting. To address this, I created DockerLinux, a free game for learning Linux and Docker in a simulated terminal.

In the game, you are tasked with investigating a simulated production server when a problem arises. You are notified of an incident via a ticket, such as SEV-1 · INC-101 · The shop is down. You log into a pretend server with real commands like docker ps -a, docker logs, curl, and docker network inspect to troubleshoot the issue. The game includes runbook.md with hints and a postmortem after you fix the problem, detailing the time taken, root cause, and lesson learned.

There are six incidents in total, each unlocked by finishing the corresponding level that teaches a specific skill. For example, to fix the "The shop is down" incident, you need to understand container lifecycle management, which is covered in Level 5. The incidents progress from container lifecycle to networking, building images, using volumes, Docker Compose, and more.

The game runs entirely in the browser, simulating a shell, virtual file system, and Docker daemon. Each incident is represented as data, including an ID, code, title, severity, required level for unlocking, SLAMs (service level agreement minutes), setup, tests, hints, and solutions. When you fix an incident, the game replays the solution against a fresh broken server to ensure the fix works.

The game was designed to help developers practice the real job of reading logs, identifying the problem, and fixing it under pressure. It provides a safe environment to learn and develop these essential skills before encountering real-life outages.