Kubernetes · course, labs and mock exams
CKAD course and practice labs
Certified Kubernetes Application Developer: 16 modules of lessons and demos, 304 tasks from 132 lab templates on real kubeadm clusters, and 15 timed mock exams.
- Modules
- 16
- Lessons
- 79
- Demos
- 18
- Labs
- 132
- Tasks
- 304
- Mock exams
- 15
The Linux Foundation · 120 minute exam
What you'll learn
Every domain of the exam, weighted like the exam
All 5 domains in the public curriculum. 5 of them already have labs.
Application Environment, Configuration and Security
25%of the exam
8 competencies · 25 labs
Application Design and Build
20%of the exam
4 competencies · 19 labs
Application Deployment
20%of the exam
4 competencies · 20 labs
Services and Networking
20%of the exam
3 competencies · 31 labs
Application Observability and Maintenance
15%of the exam
5 competencies · 37 labs
What's included
A full course, not just a question bank
79 lessons
Short lessons for the exam competencies, with commands to copy and links to the official docs.
18 demos
Walk through a task step by step before you try it yourself.
132 hands-on labs
304 tasks with their variants, on a kubeadm cluster, graded check by check.
16 module quizzes
127 original questions, with an explanation for every answer.
15 mock exams
10 full mocks and 5 domain mocks. Pass mark 66%.
Certificate of completion
From The Cloud Panda when you finish the course. It is a course certificate, not one from The Linux Foundation.
Curriculum
16 modules
79 lessons and 18 demos: 18 h 24 min of reading and demos, about 28.5 hours with the practice. The lessons open with a free account.
01Core concepts and exam setup
What the CKAD measures, how Kubernetes objects are shaped, and the kubectl habits (generators, dry runs, contexts, namespaces) that every later module relies on. Domain: Application Design and Build.
- lessonWhat the CKAD tests, and how this course maps to it9 min
- lessonKubernetes objects, YAML and kubectl explain11 min
- lessonGenerate, edit, apply: kubectl at exam speed13 min
- lessonNamespaces, contexts and working on the right cluster9 min
- demoDemo: deploy, expose and check an app in five minutes14 min
- lessonThe exam environment and the allowed documentation10 min
02Pods and multi-container patterns
The pod as the unit of deployment: its life cycle, commands and environment, and the init, sidecar, ambassador and adapter patterns for putting more than one container in it. Domain: Application Design and Build.
- lessonPod anatomy and life cycle11 min
- lessonCommands, arguments and environment variables10 min
- lessonInit containers: prepare before the app starts10 min
- lessonSidecar containers11 min
- lessonAmbassador and adapter patterns9 min
- demoDemo: add a log sidecar to a Deployment and a Job13 min
03Container images: define, build and modify
How images are built from layers, how to write and fix Dockerfiles (including multi-stage builds), how to build, tag, push and export images with Docker or Podman, and how pods pull and pin them. Domain: Application Design and Build.
- lessonImages, layers, tags and registries10 min
- lessonWriting a Dockerfile12 min
- lessonMulti-stage builds and smaller, safer images11 min
- lessonBuild, tag, push and export images with Docker or Podman11 min
- lessonUsing images in pods: pull policy, digests and private registries10 min
- demoDemo: containerize an app, fix its Dockerfile and run it in the cluster15 min
04Workload resources: choosing the right controller
Deployments, DaemonSets, StatefulSets, Jobs and CronJobs: what each controller guarantees, how to create and tune them quickly, and how to pick the right one for a task. Domain: Application Design and Build.
- lessonChoosing the right workload resource10 min
- lessonDeployments and ReplicaSets11 min
- lessonDaemonSets: one pod per node9 min
- lessonJobs: run to completion12 min
- lessonCronJobs: Jobs on a schedule11 min
- lessonStatefulSets: stable identity and per-replica storage10 min
- demoDemo: pick the right controller for three workloads14 min
05Volumes: ephemeral and persistent storage
How containers keep and share files: emptyDir and hostPath, projected configuration volumes, PersistentVolumeClaims, StorageClasses and access modes, and generic ephemeral volumes. Domain: Application Design and Build.
- lessonEphemeral volumes: emptyDir and hostPath11 min
- lessonConfigMap, Secret, downward API and projected volumes10 min
- lessonPersistentVolumes and PersistentVolumeClaims12 min
- lessonStorageClasses, access modes and generic ephemeral volumes11 min
- demoDemo: make an app's data survive pod replacement13 min
06Rolling updates, rollbacks and deployment strategies
How a Deployment rolls out a new version, how to tune, pause, recover and roll back a rollout, and how to build blue/green and canary releases from plain Deployments, labels and Services. Domain: Application Deployment.
- lessonHow a rolling update works12 min
- lessonRollout history and rollbacks10 min
- lessonPausing rollouts and recovering stuck ones11 min
- lessonDeployment strategies: recreate, rolling, blue/green and canary9 min
- lessonBlue/green deployments with a Service switch11 min
- lessonCanary deployments with replica ratios11 min
- demoDemo: canary, promote, then roll back a bad release15 min
07Packaging and configuration: Helm and Kustomize
Install, configure, upgrade and roll back existing Helm charts, inspect what they render, and customise plain manifests per environment with Kustomize bases, overlays, patches and generators. Domain: Application Deployment.
- lessonHelm concepts: charts, releases and repositories10 min
- lessonFind and install charts from a repository11 min
- lessonConfigure a chart with values11 min
- lessonUpgrade, roll back and uninstall releases11 min
- demoDemo: a release from install to rollback14 min
- lessonKustomize basics: a kustomization.yaml and kubectl -k11 min
- lessonBases, overlays and patches12 min
- lessonConfigMap and Secret generators10 min
- demoDemo: one base, two environments, and a broken overlay14 min
08Health checks, logs and monitoring
Liveness, readiness and startup probes and clean shutdowns; reading container logs; and the built-in kubectl tools for watching applications: events, describe, top, watch and precise output formats. Domain: Application Observability and Maintenance.
- lessonLiveness, readiness and startup probes12 min
- lessonTuning probes, and startup probes for slow apps11 min
- lessonGraceful shutdown: preStop hooks and grace periods9 min
- lessonContainer logs11 min
- lessonEvents and describe: what Kubernetes did to your app10 min
- lessonMonitoring apps with built-in kubectl tools12 min
- demoDemo: watch liveness and readiness probes at work14 min
09Debugging applications and API deprecations
A repeatable method for broken pods, the exec and debug tools for looking inside them, and how to keep manifests working as Kubernetes deprecates and removes API versions. Domain: Application Observability and Maintenance.
- lessonA method for debugging pods12 min
- lessonLooking inside: exec, kubectl debug, port-forward and cp11 min
- demoDemo: three broken Deployments, one method15 min
- lessonAPI versions and the deprecation policy10 min
- lessonFinding served versions and migrating manifests11 min
- demoDemo: make a folder of old manifests apply cleanly13 min
10Configuration: ConfigMaps and Secrets
Keep configuration and credentials out of images: create ConfigMaps and Secrets from literals, files and env files, consume them as variables or files, understand update behaviour, and use the built-in Secret types. Domain: Application Environment, Configuration and Security.
- lessonConfigMaps11 min
- lessonSecrets: create and consume12 min
- lessonConsuming configuration: env, envFrom, volumes and updates11 min
- lessonSecret types and handling Secrets safely9 min
- demoDemo: wire config and credentials into a Deployment, then change them13 min
11ServiceAccounts, RBAC and admission control
How every API request is authenticated, authorised and admitted; giving applications their own ServiceAccount with just the permissions they need; and recognising and fixing requests that admission control rejects. Domain: Application Environment, Configuration and Security.
- lessonAuthentication, authorization and admission11 min
- lessonServiceAccounts: identities for applications11 min
- lessonRBAC: Roles, ClusterRoles and bindings12 min
- lessonAdmission control from the developer's side11 min
- demoDemo: give an app exactly the API access it needs14 min
12Resource requirements, quotas and application security
Requests, limits and QoS; namespace LimitRanges and ResourceQuotas; and running containers safely with security contexts, Linux capabilities and the Pod Security Standards. Domain: Application Environment, Configuration and Security.
- lessonRequests and limits12 min
- lessonQoS classes, OOM kills and eviction10 min
- lessonLimitRanges: defaults and bounds per container10 min
- lessonResourceQuotas: a budget per namespace11 min
- lessonSecurity contexts: users, groups and filesystems12 min
- lessonCapabilities, privilege escalation and privileged containers10 min
- lessonPod Security Standards and Pod Security Admission11 min
- demoDemo: deploy into a namespace with a quota and the restricted level15 min
13Extending Kubernetes: CRDs and operators
Discover the resource types a cluster adds beyond the built-ins, read and write custom resources, define a simple CustomResourceDefinition, and understand how operators turn custom resources into running applications. Domain: Application Environment, Configuration and Security.
- lessonHow Kubernetes is extended, and how to discover it10 min
- lessonWriting a CustomResourceDefinition12 min
- lessonWorking with custom resources: status, RBAC and troubleshooting10 min
- lessonOperators: controllers that run applications for you11 min
- demoDemo: discover a custom resource, create one and fix it13 min
14Services and Ingress
Give pods a stable address with Services, choose the right Service type, find services by DNS, troubleshoot a Service that doesn't answer, and expose HTTP applications with Ingress rules, hosts, paths and TLS. Domain: Services and Networking.
- lessonServices: stable addresses for changing pods11 min
- lessonService types: ClusterIP, NodePort, LoadBalancer, ExternalName, headless11 min
- lessonFinding services: DNS names and environment variables10 min
- lessonTroubleshooting a Service that doesn't answer12 min
- lessonIngress: HTTP routing into the cluster11 min
- lessonIngress hosts, path types, default backends and TLS11 min
- demoDemo: expose two apps through Services and one Ingress14 min
15NetworkPolicies
Control which pods may talk to which: how NetworkPolicies select pods, combine rules and namespaces, default-deny ingress and egress without breaking DNS, and how to debug traffic a policy blocks. Domain: Services and Networking.
- lessonNetworkPolicy basics: isolation and allow rules12 min
- lessonSelectors, namespaces and ports: getting rules exactly right11 min
- lessonDefault deny, egress rules and keeping DNS working11 min
- demoDemo: lock down a three-tier app, one flow at a time15 min
16Exam strategy and mock exams
Turn knowledge into points: a per-task routine, time management across two hours, a drill sheet of the fastest commands for every domain, finding YAML in the docs, and how to use the timed mock exams. Domain: Application Design and Build.
- lessonA routine for every task, and for the whole two hours10 min
- lessonSpeed drills: the fastest command for every domain12 min
- lessonFinding the right YAML in the allowed docs, fast9 min
- demoDemo: the first twenty minutes of a mock exam15 min
How a lab works
The same loop as a hands-on exam
A task, a live cluster, and checks that read the cluster.
We set the scene
A setup script puts a live environment into a real state, often a broken one: a broken workload, a missing permission, a Service with no endpoints. You get the task and a terminal.
You fix it
With the real tools and the official documentation, exactly as you will in the exam. Nothing is simulated and there are no multiple-choice shortcuts.
Checks grade it
Each lab runs checks against the live environment and shows which passed. There is no answer text to match: if a check fails, you know exactly what is left.
Lab library
Every lab, by exam domain
304 hands-on tasks across 5 domains, about 41 hours of practice.
- Application Environment, Configuration and Security25%
- Application Design and Build20%
- Application Deployment20%
- Services and Networking20%
- Application Observability and Maintenance15%
Application Environment, Configuration and Security 25% of the exam · 25 labs
Discover and use resources that extend Kubernetes (CRD, Operators)
- E1Discover a CRD and create a custom resource (2 variants)7 min
- E2Work with custom resources and their validation rules (2 variants)8 min
- E21Create a custom resource from its served version (2 variants)8 min
Understand authentication, authorization and admission control
- E3Least-privilege RBAC for an application's ServiceAccount (2 variants)8 min
- E4One ClusterRole, granted in selected namespaces only (2 variants)8 min
- E5Fix an application that the API server answers with 403 Forbidden (3 variants)8 min
- E6Pods rejected at admission by a LimitRange (3 variants)9 min
- E22Role and RoleBinding for a user (2 variants)8 min
Understand requests, limits, quotas
- E7ResourceQuotas: compute and object counts (3 variants)9 min
- E8A ResourceQuota blocks a rolling update (3 variants)8 min
Understand ConfigMaps
- E11ConfigMaps as environment variables (2 variants)7 min
- E12ConfigMap volumes: stale subPath files and selected keys (2 variants)9 min
- E13Immutable ConfigMaps and versioned configuration (2 variants)8 min
- E23An explicit env var overrides envFrom (2 variants)8 min
Define resource requirements
- E9Requests, limits and the resulting QoS class (2 variants)8 min
- E10Pending pods and a pod's effective resource request (2 variants)9 min
Create & consume Secrets
- E14Create a Secret and consume it as env and as files (2 variants)8 min
- E15TLS and registry Secrets, pulled in through a ServiceAccount (2 variants)9 min
- E16One projected volume from a Secret, a ConfigMap and the Downward API (2 variants)8 min
- E24Rotate a Secret value and restart the workload (2 variants)8 min
Understand ServiceAccounts
- E17A dedicated ServiceAccount without an automatic API token (2 variants)8 min
- E18ServiceAccount tokens with an audience and an expiry (2 variants)9 min
Understand Application Security (SecurityContexts, Capabilities, etc.)
- E19SecurityContext: non-root user, fsGroup and a read-only root filesystem (2 variants)9 min
- E20Pod Security Admission: make a workload pass the restricted standard (2 variants)10 min
- E25Drop capabilities and set the runtime default seccomp profile (2 variants)9 min
Application Design and Build 20% of the exam · 19 labs
Define, build and modify container images
- D1Containerize an app with a Dockerfile (3 variants)8 min
- D2Repair a multi-stage Dockerfile (3 variants)8 min
- D3Pin a Deployment's images by tag and digest (3 variants)7 min
- D5Inspect an image, then override its CMD but keep its ENTRYPOINT (3 variants)8 min
Choose and use the right workload resource (Deployment, DaemonSet, CronJob, etc.)
- D4One agent per node: Deployment to DaemonSet (3 variants)8 min
- D6Replace a sleep loop with a CronJob (3 variants)8 min
- D7Fan out work with an Indexed Job (3 variants)8 min
- D8A one-off task that restarts for ever: run it as a Job (3 variants)7 min
- D17Parallel Job without completion indexes (2 variants)8 min
Understand multi-container Pod design patterns (e.g. sidecar, init and others)
- D9Init containers: wait for a dependency, then render content (3 variants)8 min
- D10A Job that never completes: make its log shipper a native sidecar (3 variants)8 min
- D11Adapter container: expose an app's status as metrics (3 variants)8 min
- D12Ambassador container: move the backend without touching the app (3 variants)9 min
- D19Share the pod process namespace with a sidecar (2 variants)8 min
Utilize persistent and ephemeral volumes
- D13Read-only root filesystem: give nginx emptyDir scratch space (3 variants)8 min
- D14Claim a PersistentVolume that already holds data (3 variants)8 min
- D15Projected volume: an audience-bound token instead of the default one (3 variants)8 min
- D16Config and credentials as files: repair ConfigMap and Secret volumes (3 variants)9 min
- D18Downward API volume for labels and a CPU limit (2 variants)9 min
Application Deployment 20% of the exam · 20 labs
Use Kubernetes primitives to implement common deployment strategies (e.g. blue/green or canary)
- P1Blue/green: switch a Service to the new version (3 variants)7 min
- P2Canary: split traffic by replica count (3 variants)7 min
- P3Canary that gets no traffic (3 variants)8 min
- P14Promote a canary without losing capacity (2 variants)7 min
- P17Recreate strategy for a single-replica app (2 variants)7 min
Understand Deployments and how to perform rolling updates
- P4From Recreate to a tuned rolling update (3 variants)7 min
- P5Roll back to a known-good revision (3 variants)6 min
- P6Pause a rollout to batch several changes (3 variants)6 min
- P7Finish a stuck rollout (3 variants)8 min
- P18Rolling update with percentage surge and a ready window (2 variants)8 min
Use the Helm package manager to deploy existing packages
- P8Helm: install from a repository with overrides (3 variants)7 min
- P9Helm: render a chart, then install it with a values file (3 variants)7 min
- P10Helm: upgrade, roll back and uninstall releases (3 variants)8 min
- P15Helm: upgrade a release to an approved chart version (3 variants)8 min
- P20Helm upgrade --install with --set (2 variants)8 min
Kustomize
- P11Kustomize: one base, two overlays (3 variants)9 min
- P12Kustomize: a generated ConfigMap that rolls the Deployment (3 variants)7 min
- P13Kustomize: repair a broken overlay (3 variants)8 min
- P16Kustomize: labels that spare the selectors, a Secret generator and a JSON patch (3 variants)9 min
- P19Kustomize name suffix and shared annotations (2 variants)9 min
Services and Networking 20% of the exam · 31 labs
Demonstrate basic understanding of NetworkPolicies
- N8Allow one client into a server (2 variants)10 min
- N9Allow a namespace and a pod label together (2 variants)12 min
- N10Default-deny egress and keep DNS (2 variants)12 min
- N23Egress only to a Service in another namespace (2 variants)12 min
- N24Default-deny ingress, then allow one namespace (2 variants)12 min
- N25Open the port a NetworkPolicy actually needs (2 variants)8 min
- N26One policy with ingress and egress rules (2 variants)12 min
- N27NetworkPolicy: default deny, then allow the frontend (2 variants)8 min
- N28NetworkPolicy: admit a monitoring namespace by label (2 variants)8 min
- N29NetworkPolicy: an egress allow list with DNS (2 variants)9 min
- N30NetworkPolicy: label a new app to fit the existing policies (2 variants)8 min
- N31NetworkPolicy: fix a policy that doesn't do its job (3 variants)7 min
Provide and troubleshoot access to applications via services
- N1Expose a Deployment and fetch it through the Service (2 variants)8 min
- N2Create a Service with the right selector and targetPort (2 variants)8 min
- N3NodePort, headless and ExternalName Services (2 variants)10 min
- N4Reach a Service by its cross-namespace DNS name (2 variants)8 min
- N5Repair a Service with no endpoints (2 variants)10 min
- N11Service with an HTTP port and a metrics port (2 variants)8 min
- N12Point targetPort at a named container port (2 variants)7 min
- N13ClientIP session affinity on a Service (2 variants)7 min
- N14Service without a selector, backed by an EndpointSlice (2 variants)10 min
- N15Change a Service port from UDP to TCP (2 variants)7 min
- N16Switch a Service to NodePort without changing its cluster IP (2 variants)8 min
- N17Add a second port without dropping the first (2 variants)7 min
- N18Report a Service cluster IP, port, and endpoint count (2 variants)6 min
Use Ingress rules to expose applications
- N6Ingress host and path rules (2 variants)10 min
- N7Ingress TLS, exact path and default backend (2 variants)12 min
- N19One Ingress, two hosts (2 variants)8 min
- N20Exact and Prefix paths on one host (2 variants)8 min
- N21Add a path to an existing Ingress (2 variants)7 min
- N22Ingress default backend plus one host rule (2 variants)8 min
Application Observability and Maintenance 15% of the exam · 37 labs
Understand API deprecations
- O1Migrate removed Deployment and Ingress APIs (2 variants)10 min
- O2Record the preferred API version and apply it (2 variants)8 min
- O12Migrate a CronJob off batch/v1beta1 (2 variants)8 min
- O13Migrate a PodDisruptionBudget to policy/v1 (2 variants)7 min
- O14Migrate a HorizontalPodAutoscaler to autoscaling/v2 (2 variants)8 min
- O15Migrate a Role and RoleBinding to rbac v1 (2 variants)8 min
- O28Removed and deprecated APIs: HPA, PDB and Endpoints (2 variants)8 min
Implement probes and health checks
- O3Fix liveness and readiness, then drain on shutdown (2 variants)12 min
- O4Set startup, liveness and readiness timing (2 variants)10 min
- O5Add a startup probe for a slow command (2 variants)10 min
- O16TCP liveness and an exec readiness probe (2 variants)8 min
- O17Point the readiness probe at a path nginx serves (2 variants)8 min
- O18Drain with a preStop sleep and a longer grace period (2 variants)8 min
- O19HTTP startup probe for a slow nginx start (2 variants)10 min
- O29Add readiness and liveness probes (3 variants)7 min
- O30Fix misconfigured health checks (3 variants)7 min
- O31Startup probe for a slow-starting app (2 variants)7 min
Use built-in CLI tools to monitor Kubernetes applications
- O6Report phase, waiting reason and ready pods (2 variants)8 min
- O7Read the scheduling event and fix the pod (2 variants)8 min
- O20Report a pod IP, host IP, and node name (2 variants)6 min
- O21Report Deployment availability (2 variants)6 min
- O22Report a completed pod phase and reason (2 variants)6 min
- O23Report QoS class from the pod status (2 variants)7 min
- O32Find the heaviest workload with kubectl top (2 variants)6 min
- O33Pod health report with kubectl get (2 variants)7 min
Utilize container logs
- O8Ship a log file with a sidecar (2 variants)10 min
- O9Read the init container logs and fix the write (2 variants)8 min
- O24Save current logs and the previous container logs (2 variants)8 min
- O25Save logs from each container in a pod (2 variants)7 min
- O34Logs from multi-container pods and label selectors (2 variants)6 min
Debugging in Kubernetes
- O10Mount the ConfigMap the crashing app reads (2 variants)8 min
- O11Save a port-forward response and an exec'd file (2 variants)8 min
- O26Fix an image tag that does not exist (2 variants)7 min
- O27Add the Secret key a pod is waiting on (2 variants)8 min
- O35Why isn't the pod running? Read the status and fix it (3 variants)7 min
- O36Debug a shell-less container with kubectl debug (2 variants)8 min
- O37Let the events explain a stuck pod (3 variants)7 min
Mock exams
15 timed mock exams
Pass mark 66%. A task appears in only one mock.
| Mock | Kind | Tasks | Time |
|---|---|---|---|
| CKAD Domain Mock: Application Deployment | One domain | 10 | 120 min |
| CKAD Domain Mock: Application Design and Build | One domain | 10 | 120 min |
| CKAD Domain Mock: Application Environment, Configuration and Security | One domain | 10 | 120 min |
| CKAD Domain Mock: Services and Networking | One domain | 10 | 120 min |
| CKAD Domain Mock: Application Observability and Maintenance | One domain | 11 | 120 min |
| CKAD Full Mock 01 | Full exam | 20 | 120 min |
| CKAD Full Mock 02 | Full exam | 20 | 120 min |
| CKAD Full Mock 03 | Full exam | 20 | 120 min |
| CKAD Full Mock 04 | Full exam | 20 | 120 min |
| CKAD Full Mock 05 | Full exam | 20 | 120 min |
| CKAD Full Mock 06 | Full exam | 20 | 120 min |
| CKAD Full Mock 07 | Full exam | 20 | 120 min |
| CKAD Full Mock 08 | Full exam | 20 | 120 min |
| CKAD Full Mock 09 | Full exam | 20 | 120 min |
| CKAD Full Mock 10 | Full exam | 20 | 120 min |
Start the CKAD course
A free account opens the lessons. The free starter labs on the CKA stay the CKA labs.