L01 — Architecture

What runs inside a Kubernetes cluster, and where. Once this is clear, every other level is “now do something with those components”.

What you’ll understand after this level

• The difference between control plane and worker nodes
• What every component in the control plane does (kube-apiserver, etcd, kube-scheduler, kube-controller-manager, cloud-controller-manager)
• What runs on every node (kubelet, kube-proxy, container runtime)
• How a request flows from kubectl apply to a running pod
• How multi-master HA topology works
• What a namespace is and how it differs from a regular cluster

Notes in this level

|| Note | Status | What’s in it | |------|--------|--------------| || Setting up a Cluster | ✅ | kubeadm, the-hard-way, managed clusters, cloud-native bootstrap patterns | || High Availability | ✅ | etcd quorum, multi-master, control-plane failure modes, stacked vs external | || Namespaces | ✅ | Namespace model, default namespace, resource limits, namespace deletion lifecycle | || Local Deployment | ✅ | k3d, kind, minikube comparison, dev cluster patterns | || Need for swapoff | ✅ | Why kubelet requires swap disabled, cgroup memory pressure, swap vs memory limits | || What Happens When… | ✅ | End-to-end trace: kubectl apply → etcd → scheduler → kubelet → container runtime |

Suggested reading order

1. Setting up a Cluster — get a cluster running first
2. Namespaces — the first thing to know to organize anything
3. High Availability — what “production” means in k8s
4. What Happens When… — tie it all together with a request trace
5. Local Deployment — after you understand the components, set up your dev environment

Where to go next

→ L02 — Objects: now that you know the components, learn the data model they manipulate.