Beginner

Exam Overview

Everything you need to know about the Certified Kubernetes Administrator (CKA) exam — format, domains, scoring, and the AI-specific skills that make Kubernetes essential for modern ML infrastructure.

What Is the CKA Exam?

The Certified Kubernetes Administrator (CKA) is a performance-based exam administered by the Cloud Native Computing Foundation (CNCF) and the Linux Foundation. Unlike multiple-choice exams, you solve real tasks in a live Kubernetes environment using the command line.

Exam Details at a Glance

  • Format: 15-20 performance-based tasks in a live terminal
  • Duration: 2 hours
  • Passing score: 66%
  • Cost: $395 (includes one free retake)
  • Validity: 3 years
  • Proctoring: Online, proctored via webcam
  • Open book: You can access the official Kubernetes documentation during the exam
  • Kubernetes version: The exam runs on the latest stable release

CKA Exam Domains

The CKA covers five primary domains. Understanding their weights helps you prioritize your study time.

1. Cluster Architecture, Installation & Configuration (25%)

RBAC, kubeadm cluster setup, etcd backup/restore, cluster upgrades. For AI workloads, this includes configuring nodes with GPU drivers and container runtimes that support GPU passthrough.

2. Workloads & Scheduling (15%)

Deployments, DaemonSets, ConfigMaps, Secrets, resource limits, scheduling. For AI, this is where GPU resource requests, node affinity, taints, and tolerations come in — critical for directing ML workloads to GPU nodes.

3. Services & Networking (20%)

Services (ClusterIP, NodePort, LoadBalancer), Ingress, NetworkPolicies, DNS. For AI, you need to expose model serving endpoints via Ingress and secure inter-service communication for distributed training.

4. Storage (10%)

PersistentVolumes, PersistentVolumeClaims, StorageClasses, volume modes. For AI, storage is critical: large datasets, model checkpoints, and training artifacts all need persistent, high-performance storage.

5. Troubleshooting (30%)

Debug cluster components, application failures, networking issues. For AI workloads, this includes diagnosing GPU allocation failures, OOM kills on large models, and stuck training jobs.

💡
AI focus: While the CKA itself is not AI-specific, every domain has AI/ML relevance. This course teaches core CKA concepts through the lens of running ML workloads — so you learn both Kubernetes administration AND how to support data science teams.

Why Kubernetes for AI?

Kubernetes has become the de facto platform for ML infrastructure. Here is why:

  • GPU orchestration — Schedule and manage GPU resources across a cluster
  • Scalable training — Run distributed training across multiple nodes
  • Model serving — Deploy models as scalable microservices with autoscaling
  • ML pipelines — Orchestrate end-to-end ML workflows with Kubeflow, Argo, or Airflow on K8s
  • Resource isolation — Namespaces and resource quotas keep ML experiments from starving other workloads
  • Reproducibility — Containers ensure consistent environments from development to production

Prerequisites

Before starting this course, you should have:

  • Basic Linux command line skills (navigating directories, editing files, running commands)
  • Familiarity with containers (Docker basics: images, containers, Dockerfiles)
  • Understanding of YAML syntax (indentation, key-value pairs, lists)
  • Basic networking concepts (IP addresses, ports, DNS)

No prior Kubernetes experience is required — we start from the fundamentals and build up to AI-specific configurations.

Study Plan

2-Week Intensive Plan

  • Week 1: Lessons 1-4 (Exam overview, Core concepts, GPU scheduling, ML workloads). Practice with a local cluster (minikube or kind).
  • Week 2: Lesson 5 (Networking & Storage), Practice exam, Best practices. Take 2-3 practice exams. Focus on speed — the CKA is time-pressured.

4-Week Comfortable Plan

  • Weeks 1-2: One lesson per day. Build a multi-node cluster and practice each concept hands-on.
  • Week 3: Practice exam + review weak areas. Practice kubectl speed drills.
  • Week 4: Full practice exams under timed conditions. Review best practices and exam tips.
⚠
Critical: The CKA is hands-on. Reading alone will not prepare you. You MUST practice in a real cluster. Use minikube, kind, or a cloud-based lab environment. Aim for at least 20 hours of hands-on practice before exam day.