study.

CKA learning path

Understand the concepts, practice the commands, and learn to troubleshoot a cluster.

These 15 foundational lessons are a starting point, not a complete exam preparation course. Advanced labs on kubeadm, etcd recovery, Helm, Kustomize, Gateway API, and more will follow.

You need kubectl and access to your own practice Kubernetes cluster. Each lab is independent. Check your current context first, then run the commands in your own terminal.

Learn with progress and quizzes

Essential concepts before you start

Control plane and workers

The control plane handles API access, state storage, scheduling, and reconciliation. Worker kubelets and runtimes execute assigned Pods. This distinction separates investigation of failed API requests from image pulls after node placement.

Authentication, authorization, and admission

Authentication establishes identity; authorization decides whether an action is allowed. Admission can further validate or modify permitted write requests. RBAC permission does not guarantee creation succeeds; distinguish Forbidden from quota or policy rejection.

Requests, limits, and allocatable

Allocatable is node capacity available for Pods. Scheduling considers requests and placement constraints; limits constrain runtime use. 1000m CPU is one CPU, and 128Mi uses binary units. Low utilization does not imply room if requests already reserve capacity.

Layers of the network path

Separate DNS resolution, Service backend selection, forwarding to a Pod and port, and the application response. Policies may govern source egress and destination ingress. Locating the failing segment avoids restarting healthy components without evidence.

PVs, PVCs, and StorageClasses

A PVC requests storage, a PV represents it, and a StorageClass defines provisioning behavior. Binding, mounting, and file permissions are separate stages. A Bound claim does not automatically prove the app can write files.

Diagnosing with state, events, and logs

State describes observations, events explain system actions, and logs record process or service behavior. Correlate the same time window and object or container. Change one factor at a time and retest the original failure to judge recovery.

Cluster Architecture

Understand cluster components, contexts, and access control.

Workloads & Scheduling

Deploy Pods and maintain their desired state.

Services & Networking

Connect services and trace network traffic.

Storage

Understand data lifecycles and volume binding.

Troubleshooting

Gather evidence and narrow down the cause of a failure.