Course 2 of 7 · The deployment/infra track

Kubernetes, from newcomer to interview-ready

The orchestrator that runs every container in this repo — the object model, the control loop, and how this codebase uses it.

Course 1 ended at the kind + Skaffold local stack: a real Kubernetes cluster that Skaffold deploys Helm charts into. This course opens that box. What are the objects Skaffold creates — Pods, Deployments, Services, StatefulSets, autoscalers — and how does the cluster keep them alive? This is the biggest course in the track, aligned to the CKA/CKAD syllabus. Three parts, built one at a time. Lessons are short — one win each. Read them in order.

How to use this Do one lesson, take its quiz from memory (no peeking), then skim the matching row of the cheat sheet. Come back a day later and re-take it — spacing beats cramming. Your seventh course; you know the rhythm. Stuck? Ask me — I'm your teacher, not just the author.
One honest caveat up front This repo almost never hand-writes Kubernetes YAML — it generates every object from Helm templates. So each lesson teaches the object (what the cert and interviews test) and points at the template that produces it. The templating mechanics themselves are Course 3 (Helm).

Part 1 — Core objects & the control plane available now

The mental model everything else hangs off: desired state, the control loop, and the workload hierarchy.

1 · What Kubernetes is & the control loop

Declarative desired-state and the reconciliation loop that makes it real.

ready

2 · Cluster architecture

Control plane (api-server, etcd, scheduler, controllers) vs nodes (kubelet, kube-proxy).

ready

3 · Pods — the atom

One-or-more containers sharing network + storage; ephemeral by design.

ready

4 · Deployments & ReplicaSets

Desired replica count, self-healing, and the zero-downtime rolling update.

ready

Part 2 — Workloads, config & networking available now

Connecting and configuring workloads — and when a service needs a stable identity.

5 · Services & cluster DNS

ClusterIP / headless / NodePort / LoadBalancer, and name-based discovery.

ready

6 · ConfigMaps & Secrets

Injecting config and credentials — and the repo's SOPS-encrypted secrets.

ready

7 · StatefulSets vs Deployments

Stable identity + ordered, persistent pods — why conversationmgmt is one.

ready

8 · Probes & resources

liveness / readiness / startup, and requests vs limits.

ready

Part 3 — Scaling, scheduling & reliability available now

Production behaviour: autoscaling, where pods land, and surviving disruption.

9 · Autoscaling: HPA / VPA / KEDA

The repo's ScaledObjects — event-driven scaling, even to zero.

ready

10 · Scheduling

Labels/selectors, node affinity, taints & tolerations, spot nodes.

ready

11 · Disruptions & the pod lifecycle

PodDisruptionBudgets, graceful termination, rollout safety.

ready

12 · Namespaces, RBAC & ServiceAccounts

The {env}-{org}-backend namespaces and who's allowed to do what.

ready

Reference shelf

Cheat sheet

Dense revision sheet + kubectl + interview one-liners.

Glossary

The canonical vocabulary, opinionated.

Repo K8s map

Ground truth: which templates make which objects.

Resources

kubernetes.io, KEDA, the CKA/CKAD certs.
All 12 lessons are built — the Kubernetes course is complete. From the control-loop model through the object hierarchy to production autoscaling, scheduling and RBAC, every object anchored to this repo's libs/util chart and aligned to CKA/CKAD. Next: Course 3 (Helm) explains the templating engine behind it all. Or — eight courses' worth of lessons, still no retention check — ask me to run a mock interview and I'll record where you're solid.