Deployment to ReplicaSet to Pod

A Deployment is not the thing that directly creates your Pods - there is a hidden middle layer worth understanding. A Deployment owns one or more ReplicaSets, and each ReplicaSet is the object that actually owns and maintains a set of Pods matching a label selector, making sure exactly the requested number of copies exist at all times. When you change the Pod template inside a Deployment (for example, updating the container image to a new version), Kubernetes does not edit the existing Pods in place - it creates a brand-new ReplicaSet with the new template, and gradually shifts Pod count from the old ReplicaSet to the new one. This is why, if you run kubectl get rs, you will often see old, now-empty ReplicaSets left behind - they are kept around to make instant rollback possible. The Deployment's selector field is the label expression that decides which Pods a ReplicaSet considers "mine," and it is treated as an identity contract that should essentially never change after creation, because changing it can orphan Pods or cause two controllers to fight over ownership. Two more fields shape every release: the plain replica count expresses your steady-state availability intent ("how many copies should exist when nothing is changing"), while maxSurge and maxUnavailable together define the rollout capacity envelope - how many extra Pods may temporarily exist above the target, and how many may temporarily be missing below it, while a rollout is in progress.

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 selector expression is written too loosely and accidentally also matches Pods that belong to a completely unrelated Deployment. Both controllers now believe they own the same Pods, leading to ownership conflicts, Pods being unexpectedly deleted or recreated, and confusing scale events that nobody triggered. The fix is to always use stable, specific labels for identity (like app: checkout) and keep transient release metadata, such as a version or timestamp, out of the selector entirely - selectors should describe what this is, not which version this happens to be right now.

Trace the ownership chain

Use kubectl get deploy web -o jsonpath='{.metadata.generation}{" "}{.status.observedGeneration}', then list ReplicaSets by label and compare their pod-template-hash. kubectl describe deploy web reveals desired, updated, available and unavailable replicas plus rollout conditions. Never edit a ReplicaSet or Pod expecting permanence; the Deployment template is the durable intent.

Scenario: A Deployment reports three desired and only two available. One new Pod is Pending because surge capacity exceeds quota. The image and probe are fine; FailedCreate and quota events identify the actual constraint.
Warning: Changing selector labels during a migration can orphan old Pods or overlap another controller. Treat selectors as stable identity and put version labels only in the template.

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 rolling-update-hpa-scaling. Open /labs/kubernetes, choose rolling-update-hpa-scaling, 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.