KCNA Tests Which Kubernetes Component Does What — Not How to Run kubectl
KCNA is a 60-question, multiple-choice foundation exam. Nearly half of it is Kubernetes fundamentals, so knowing exactly which component, object or project owns a job beats hands-on command recall.
Check Your Readiness →Most candidates understand Kubernetes and Cloud Native Associate (KCNA) concepts — and still fail. This exam tests how you apply knowledge under pressure.
The current four-domain outline replaced the 2021 five-domain version, and observability now sits inside Cloud Native Architecture. Questions are short identification items: pick the component, object, project or standard that fits a one- or two-line scenario.
Add a toleration to the GPU Pods so they are attracted there
Taint the GPU node so ordinary Pods are repelled, then give GPU Pods a matching toleration plus node affinity or a nodeSelector. Tolerations only permit scheduling onto a tainted node; they never pull a Pod toward it.
Run kubectl apply from a CI pipeline after each merge
This is GitOps: an in-cluster agent such as Argo CD or Flux pulls the desired state from Git and continuously reconciles the cluster. A push-based CI job applies once and does not detect or fix later drift.
Put each workload in its own Service with no selector
Apply a default-deny NetworkPolicy, then allow ingress to the API Pods from Pods labelled as frontend. NetworkPolicy only takes effect if the cluster's CNI plugin enforces it, such as Calico or Cilium.
Configure a HorizontalPodAutoscaler with minReplicas set to zero
Scale to zero is a serverless pattern, provided on Kubernetes by projects such as Knative or KEDA. The HPA scales on metrics between a minimum of one and a maximum; the Cluster Autoscaler adds or removes nodes, not Pods.
CNCF published a new curriculum on 24 November 2025: 44/28/16/12, with Cloud Native Observability folded into Architecture. Courses still showing a separate 8% observability domain underweight Container Orchestration, which rose to 28%.
kube-scheduler only picks a node by filtering and scoring. The kubelet starts containers through the CRI, and controllers in kube-controller-manager, such as the ReplicaSet controller, recreate missing Pods. Questions swap these three on purpose.
Stable names (db-0, db-1) and one volume claim per replica point to a StatefulSet. One Pod on every node, including new ones, points to a DaemonSet. A Deployment fits stateless, interchangeable replicas only.
Requests drive scheduling; limits cap usage at runtime. A container over its memory limit is OOMKilled, while one over its CPU limit is throttled, not killed. Limits never influence which node the scheduler picks.
The kubelet talks to any CRI-compliant runtime, usually containerd or CRI-O. Images built with Docker still run because they follow the OCI image format; the dockershim adapter was removed in Kubernetes 1.24.
Strengthen weak areas with exam-style practice questions and detailed explanations.
Simulate the real exam experience and assess your readiness under timed conditions.
Review key concepts, objectives, and exam topics in one place.
Get personalized explanations, learning recommendations, and instant answers.
Follow a structured learning path designed to help you prepare efficiently.
KCNA rewards knowing exactly which component owns each job. Check whether yours are sorted.