Helm, Kustomize and Release Assets

Writing every environment's YAML by hand quickly becomes unmanageable, so two popular tools exist to generate it instead - and they solve the problem in very different ways. Helm treats an application as a templated, versioned package: you fill in a values.yaml file, Helm renders full Kubernetes manifests from a chart template, and it also tracks release history so you can helm rollback an entire application's install. Kustomize takes the opposite approach: it never templates anything - instead you keep a plain, valid base set of YAML and layer small, explicit declarative patches ("overlays") on top for each environment, so a reviewer can see the exact diff between environments rather than reverse-engineering a template. As a rule of thumb, reach for Kustomize when you mainly need controlled, reviewable differences between environments of one system you own, and reach for Helm when you are packaging an application to be distributed and installed by other people through a stable values API. Whichever tool you use, always render the final YAML and inspect or diff it before applying (helm template or kubectl diff -k), pin exact chart or base versions rather than tracking a moving target, and treat the generated output itself as production code worth reviewing - not just the template that produced it.

Analogy: Kubernetes is a thermostat, not a remote control. You declare the temperature; independent controllers measure reality and keep acting. Debugging means finding which sensor, rule or actuator prevents convergence.

Read the object as evidence

kubectl get RESOURCE NAME -n NAMESPACE -o yaml
kubectl describe RESOURCE NAME -n NAMESPACE
kubectl get events -n NAMESPACE --sort-by=.metadata.creationTimestamp
kubectl explain RESOURCE.spec

Do not memorize these as a ritual. The first command exposes desired and observed state, the second connects conditions and events, the third supplies a timeline, and the fourth checks the server's schema. Compare what the controller was asked to do with what it reports doing. Then test the narrowest hypothesis.

Scenario: A Helm chart's dependency is left unpinned to a specific version, so the exact same source Git commit produces different rendered Kubernetes resources depending on when it happens to be deployed and which dependency version was resolved that day. Locking dependency versions explicitly, and storing reviewed, committed values files per environment, is what makes a deployment reproducible - without that, "redeploy the same commit" is not actually guaranteed to redeploy the same thing.

Render, diff and own generated YAML

Use helm template or helm upgrade --install --dry-run to inspect rendered resources, and kubectl diff -k overlays/prod before apply. Keep secrets out of values committed to source. In Helm, understand that a release owns rendered objects and hooks can run imperative work. In Kustomize, keep bases reusable and overlays small enough that reviewers can see the effective change.

Scenario: A production overlay replaces an entire container list to change one environment value, silently dropping probes and resources from the base. Rendered-diff review catches patch semantics that source review misses.
Tip: Pin chart and image versions, commit dependency locks, and test rendered manifests against the target API and admission policies in CI. Reproducibility includes generator inputs, not just the final Git SHA.

Production reasoning

Ask four questions: Who owns this object? What dependency must become ready next? Which controller reports the blocking condition? What evidence would disprove my current theory? This prevents symptom-driven changes. Record the context, namespace, object generation, image digest and recent rollout before mutation; a recreated Pod may erase the evidence you needed.

Warning: Running is not the same as ready, healthy, durable or correct. Kubernetes status is layered. Confirm the application-level outcome as well as the object state.
Goal: Put this model into practice in the Kubernetes lab configmap-secret-mount. Open /labs/kubernetes, choose configmap-secret-mount, predict the failure path before changing anything, then use the simulator's check command to validate the finished state.

Deliberate practice

Before the lab, write the expected object relationship and the first three commands you will run. Afterward, explain why the fix converged and name one tempting change that would only mask the symptom. Repeat using an explicit namespace and a structured output format. This prediction-observation-explanation loop is what turns command familiarity into production judgment.

45-minute investigation

  1. Map (5 min): draw the owner-to-child chain and mark every namespace, selector, identity and dependency involved.
  2. Predict (5 min): write one expected status condition, one likely event and one log or metric signal before opening the lab.
  3. Observe (10 min): collect YAML, describe output and ordered events. Do not mutate state. Record which observation disproves your first theory.
  4. Repair (15 min): make the smallest declarative correction, watch the responsible controller converge, and verify the user-facing path rather than stopping at Running.
  5. Stress (10 min): change one relevant constraint - replica count, label, readiness, resource value or placement rule - predict the outcome, observe it, then restore the known-good declaration.
Tip: Keep a short incident note with symptom, evidence, hypothesis, change, result. Across four sections this produces a reusable runbook instead of a pile of remembered commands.