Fourteen days of drills on Docker and Kubernetes, built around one real application and one real cluster that stays up the whole way through.
This is a drill manual, not a narrative course. Every listing runs against a real Docker Engine and a real local Kubernetes cluster, and every printed output block is a paste from that run — including a real crash loop caused by a liveness probe set too eager, a real registry migration that broke on the very first try, and the real VM restarts that interrupted the work more than once. Each chapter states a rule in as few words as it can, shows the rule working in a short numbered example with its captured output, and hands it straight back as problems. Every example builds one application, hiretrack, a small job-application tracker that grows a real feature every day: a SQLite file, then Postgres and Redis behind Compose, then a full Kubernetes Deployment with probes, an init container, a CI/CD pipeline, and a real container registry.
The first week covers Docker: images and layers, a fast Dockerfile, multi-stage builds, volumes and bind mounts, container networking, a real Compose stack, and a systematic method for debugging a container that will not start. The second week moves the same application onto Kubernetes: Deployments, Services, ConfigMaps and Secrets, health probes and rolling updates, init containers and paused rollouts, kubectl-based debugging of planted failures, a real CI/CD pipeline against a real registry, and a closing chapter that runs the whole book back as a timed interview drill.
The measurements are reported as they came out. A one-line source edit that should touch nothing costs a full nine-second dependency reinstall, because of Dockerfile order, not bad luck. A liveness probe with too short a delay kills a perfectly healthy container, repeatedly, until the delay is fixed. A registry-based deploy fails on the very first attempt with a pull-policy conflict this book's own earlier chapters caused. A book that only printed the clean successes would be easier to write and worth less to work through.
Your 14-Day Path
- 1. IMAGES, LAYERS AND THE CONTAINER
- 2. WRITING A DOCKERFILE THAT BUILDS FAST
- 3. MULTI-STAGE BUILDS AND SMALL IMAGES
- 4. VOLUMES, BIND MOUNTS AND DATA
- 5. CONTAINER NETWORKING
- 6. COMPOSE FOR A REAL STACK
- 7. DEBUGGING A CONTAINER THAT WON'T START
- 8. IMAGE SECURITY AND SUPPLY CHAIN
- 9. FROM COMPOSE TO KUBERNETES
- 10. DEPLOYMENTS, SERVICES AND CONFIG
- 11. HEALTH CHECKS, PROBES AND ROLLOUTS
- 12. LOGS, METRICS AND kubectl DEBUGGING
- 13. CI/CD: BUILD, PUSH, DEPLOY
- 14. THE INTERVIEW
Who It's For
Developers preparing for a DevOps or platform-engineering interview. Backend engineers whose team just adopted Kubernetes. Anyone who has run docker run a hundred times and still cannot explain what a Service actually routes to. No prior Kubernetes is assumed; comfort with a shell and a text editor is.
302 worked examples. 302 figures. 2,053 numbered problems. 350 pages. Measured with Docker Engine 29.6.1, kubectl v1.36.4, and a kind v0.33.0 cluster running Kubernetes v1.35.0.
You need Docker Desktop or an equivalent Docker Engine, and about an hour a day. Start Chapter 1.
