Introduction

Index · 02 Core Concepts →

Contents

• 1.1 Course Introduction
• 1.2 Certification

1.1 Course IntroductionCopy Anchor Link

Source: KodeKloud notes

Course approachCopy Anchor Link

• CKA is a hands-on, performance-based exam: concepts alone are not enough; lots of lab practice is needed.
• Course flow: core concepts (with analogies) → build a cluster from scratch to see how the components fit together → end-to-end validation tests and troubleshooting.
• Context: Kubernetes is the industry standard for production distributed apps; Indeed data showed Kubernetes job searches up ~173% YoY.

Lab environmentCopy Anchor Link

• Each lab matches a lesson. Tasks are challenges: inspect the environment, find the information, then apply the right command or config. Answers are auto-validated with instant feedback.
• Incremental hints and full solutions are available when you are stuck.
• The prompt is controlplane ~ ➜, so you work on the control-plane node.

kubectl get pod                    # "No resources found in default namespace."
kubectl run nginx --image=nginx    # creates a single Pod: pod/nginx created
kubectl get pod                    # NAME nginx  READY 1/1  STATUS Running

Reading kubectl describe pod (sample from the lab)Copy Anchor Link

Fields in the sample output worth recognising:

Labels:            tier=busybox
Status:            Running
IP:                10.22.0.10
Controlled By:     ReplicaSet/newpods            # owner: the Pod was created by a ReplicaSet
Containers:
  busybox:
    Container ID:  containerd://51bf...           # runtime prefix: containerd, not docker
    Image:         busybox
    Image ID:      docker.io/library/busybox@sha256:fd7d...   # short image name resolved to docker.io/library/<img>, pinned by digest
    Command:       sleep 1000
    State:         Running
    Ready:         True
    Restart Count: 0
    Mounts:
      /var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-bqjhg (ro)
Conditions:        Initialized / Ready / ContainersReady / PodScheduled = True
Volumes:
  kube-api-access-bqjhg:
    Type:                    Projected            # auto-injected ServiceAccount token volume
    TokenExpirationSeconds:  3607                 # time-bound token

• Controlled By shows the owning controller (here a ReplicaSet).
• Conditions in the sample: PodScheduled → Initialized → ContainersReady → Ready (lifecycle order).

[!NOTE] Beyond the course:

• kubectl run now creates only a Pod. The old --generator flag is gone, so use kubectl create deployment to get a Deployment.
• On current clusters the runtime is containerd/CRI-O (dockershim was removed in v1.24), so on nodes you debug containers with crictl, not docker.
• Projected, time-bound SA tokens (the kube-api-access-* volume above) have been the default since v1.21/1.22. Since v1.24, legacy token Secrets are no longer auto-created; use kubectl create token <sa> for a token on demand.
• Controlled By: use it to work out why a deleted Pod keeps coming back (its controller recreates it; delete/scale the owner instead).
• Pods get (unless automountServiceAccountToken: false) a projected kube-api-access-* volume mounted read-only at /var/run/secrets/kubernetes.io/serviceaccount. It holds the SA token, ca.crt and namespace.
• Since v1.29 a fifth condition, PodReadyToStartContainers (sandbox + networking ready), is also shown by default.

1.2 CertificationCopy Anchor Link

Source: KodeKloud notes

The CKA programCopy Anchor Link

• Run by the CNCF together with the Linux Foundation: https://www.cncf.io/certification/cka.
• Proves practical skill in designing, building and maintaining production-grade clusters. Relevant for platform, DevOps and SRE roles.
• Retake policy and exam rules change over time. Always check the Candidate Handbook on the certification site.

Exam format and rulesCopy Anchor Link

[!CAUTION] Correction: the course gives the exam length as three hours. The CKA has been 2 hours since the Sept 2020 revision, with roughly 15–20 tasks and a passing score of 66%. Confirm the current figures in the Candidate Handbook.

[!NOTE] Beyond the course: current CKA facts (verify against the handbook, because they change):

• Curriculum weights since the Feb 2025 revision: Troubleshooting 30%, Cluster Architecture/Installation & Configuration 25%, Services & Networking 20%, Workloads & Scheduling 15%, Storage 10%. Helm, Kustomize, Gateway API, CRDs/operators and CNI/CSI/CRI are explicitly in scope.
• The exam runs in a remote Linux desktop via the PSI Secure Browser. The docs are opened in the browser inside that desktop, and the allowed list is in the handbook (kubernetes.io/docs, kubernetes.io/blog and their subdomains, among others). In the terminal, copy/paste is Ctrl+Shift+C / Ctrl+Shift+V.
• Each task names the host/context to use: run the given ssh <host> (current exam) or kubectl config use-context <ctx> before starting it. Working on the wrong cluster scores zero for that task.
• The purchase includes one free retake and two killer.sh simulator sessions.

Exam strategy (course tips)Copy Anchor Link

• Practise often on real clusters: minikube, k3s, kind, or a cloud-managed cluster. Have at least one local and one cloud-managed cluster.
• Learn to find things in kubernetes.io/docs fast using the table of contents, search and anchor links.
• Time management: skip and flag slow tasks, then come back to them.
• Memorise common kubectl commands, YAML snippets, and label/selector syntax.
• Scripts are fine for practice. Check the exam rules before relying on any automation.
• Practise under time limits. Get fast at editing YAML and validating manifests.

# Core commands the course says to have in muscle memory
kubectl get <resource>
kubectl apply -f <file>.yaml
kubectl describe <resource> <name>
kubectl edit <resource> <name>
kubectl logs <pod>
kubectl exec -it <pod> -- sh

[!NOTE] Beyond the course: generate YAML instead of typing it, and validate before you apply.

Copy Text

alias k=kubectl                                   # usually pre-set in the exam env, with completion
export do="--dry-run=client -o yaml"              # e.g. k run web --image=nginx $do > pod.yaml
kubectl create deployment web --image=nginx --replicas=3 --dry-run=client -o yaml > deploy.yaml
kubectl explain pod.spec.containers --recursive   # field reference without leaving the terminal
kubectl apply -f deploy.yaml --dry-run=server     # server-side validation, nothing persisted

ResourcesCopy Anchor Link

Suggested study plan (8–12 weeks)Copy Anchor Link

[!WARNING] Exam traps / remember:

• Every task is hands-on on a live cluster. Partial knowledge with no speed does not pass, so practise under a timer.
• Only the official docs are allowed. Bookmarks and notes from outside are not.
• Read the Candidate Handbook before exam day: workspace, ID, system check and proctoring rules.

Index · 02 Core Concepts →

Reference

Based on the KodeKloud Certified Kubernetes Administrator (CKA) course notes.