This site is a learning resource for engineers who want to become more familiar with Kubernetes. It explains how Kubernetes works one layer at a time, showing the requests, manifests and source code involved so the reader can build a model they can reason with.

It is written by someone who has used Kubernetes a little, from the outside. Writing it is also the author’s way of learning more, so the pages follow what was traced and checked along the way, and say where something has not been verified.

It was motivated by Kubernetes for Haters by David J. Josephsen, and the first page, How Kubernetes works , is adapted from that article. The pages after it go further, following each thread in more detail.

Source claims are checked against kubernetes/kubernetes at v1.37.1 . Every source link on the site points at that tag.

Links marked with ↗ lead off this site. They all open in one named browser tab, external-link, so following several of them reuses the same tab instead of opening a new one each time. Put that tab beside this one with your browser’s split view (or side-by-side windows), and the source stays visible while you read.

• How Kubernetes Works: One API, Many Controllers Kubernetes is a distributed system built around a single HTTP API server. Everything else is a small, independent controller watching that API.
• What is a kubelet? The kubelet is the controller that runs on every node and turns Pod objects from kube-apiserver into running containers.