CKA (Other exams) case studies
CKA · Put your knowledge to work in context.
The Certified Kubernetes Administrator (CKA) is a vendor-neutral, proctored, performance-based certification created by the Cloud Native Computing Foundation in collaboration with The Linux Foundation. It is aimed at engineers who administer Kubernetes clusters day to day - platform, DevOps and SRE practitioners who need to prove they can operate a cluster rather than just describe one. The CKA case studies are 5 cases and about 4,346 words.
1kubeadm upgrade of a stacked-etcd HA control plane with version skew and expiring certificates
A three-node stacked-etcd control plane (cp-1, cp-2, cp-3) and two workers run Kubernetes v1.29.3, built with kubeadm, containerd as the runtime, pod CIDR 10.244.0.0/16. The platform team must reach v1.30.1 inside a 90-minute maintenance window with no API downtime.
841 words, about 4 min2A Deployment stuck one replica short: taint, requests, and a priority class that cannot help
Namespace batch runs Deployment report-gen with 6 replicas, each requesting cpu: 1500m and memory: 3Gi with no limits. There are three workers: node-a and node-b are 8 vCPU / 32Gi, node-c is 4 vCPU / 16Gi.
835 words, about 4 min3ClusterIP resolves but times out: a NetworkPolicy whose namespaceSelector matches nothing
Namespace prod runs Deployment api (3 replicas, containerPort 8080, readiness probe GET /healthz). Service api is ClusterIP on port 80 with targetPort 8080; kubectl get endpoints api -n prod lists all three pod IPs and all pods are Ready.
845 words, about 4 min4A Released PersistentVolume and a recreated PVC: restoring a StatefulSet volume without losing data
Namespace data runs StatefulSet db (3 replicas) whose volumeClaimTemplates request 20Gi with storageClassName manual . Three PVs - pv-db-0 , pv-db-1 , pv-db-2 - were hand-built: 20Gi, manual , persistentVolumeReclaimPolicy: Retain , and nodeAffinity pinning each to node-a, node-b, node-c respectively.
922 words, about 5 min5Node NotReady after reboot: kubelet running, container runtime disabled
After a scheduled OS reboot, kubectl get nodes shows node-4 NotReady and has done for 12 minutes. On the node, systemctl status kubelet reports active (running) , but crictl ps fails with connect: connection refused on unix:///run/containerd/containerd.sock , systemctl is-active containerd returns inactive , and systemctl is-enabled containerd returns disabled .
903 words, about 5 min