CKAD study plan

Create ConfigMaps and Secrets both imperatively and from manifest files, then inject them as single env vars, as envFrom blocks, and as mounted volumes

Mount a Secret as a volume inside a Pod and verify file contents and permissions from within the running container

Define CPU and memory requests and limits on a Pod and observe how the resulting QoS class changes

Create a ResourceQuota in a namespace and confirm that an over-sized Pod is rejected at admission

Attach a dedicated ServiceAccount to a Pod and set automountServiceAccountToken to false on the default account

Configure runAsUser, runAsNonRoot, readOnlyRootFilesystem, and dropped capabilities in a SecurityContext, then confirm the effect with an in-container id and mount check

Create a CustomResourceDefinition, then create and query an instance of that custom resource

Trace the request path from kubeconfig credentials through authentication, authorization, and admission, and write it out in your own words

Write a Dockerfile, build an image, and run it in-cluster with an overridden command and args

Create a Job with explicit completions and parallelism, plus a CronJob with a concurrencyPolicy and a deadline

Implement an init container and a sidecar container in one Pod and verify execution ordering from logs

Attach emptyDir and hostPath volumes to a Pod and demonstrate data sharing between two containers

Perform a rolling update on a Deployment and deliberately trigger, then repair, a failed rollout

Roll back a Deployment using both rollout undo and inspection of the revision history

Tune maxSurge and maxUnavailable and observe how scale-up and scale-down pacing changes

Scale a Deployment and inspect the generated ReplicaSets to explain revision tracking

Implement a blue/green cutover by editing a Service selector, then a canary release by splitting replicas across two Deployments

Install an existing Helm chart with a custom values file and confirm the rendered resources match the cluster state

Build Kustomize base plus dev and prod overlays from one directory tree

Render manifests with helm template and kustomize build and diff them against what is actually deployed

Expose a Deployment through ClusterIP, NodePort, and headless Services and resolve each name from a test Pod

Diagnose a Service whose selector matches no Pods and correct the endpoint list

Create an Ingress with both path-based and host-based routing and verify it against the running Ingress controller

Introduce a targetPort mismatch and explain the resulting timeout from the Service definition alone

Write NetworkPolicies that default-deny ingress and then allow traffic only from a specifically labeled namespace

Add liveness, readiness, and startup probes and verify each failure mode using kubectl describe

Stream logs from a multi-container Pod using the container flag and the previous-instance flag

Debug a CrashLoopBackOff and an ImagePullBackOff using describe, logs, and events

Identify and replace deprecated API versions flagged by kubectl warnings in existing manifests

Monitor a running workload's health using kubectl top, events, and exec-based checks

Complete two full-length timed practice exams and classify every miss by domain

Re-drill your three slowest task types until each is finished in under three minutes