Skip to content

Scoring Rubric - Kubernetes

Score 0-1

  • Don't know what Kubernetes is for
  • Haven't used kubectl

Score 2-3

  • Can explain pods, deployments, services at a high level
  • Have deployed something with kubectl apply -f
  • Need to look up every kubectl command
  • Don't understand how services route to pods

Score 4-5

  • Can create deployments, services, configmaps from memory
  • Understand labels, selectors, and how services find pods
  • Can read kubectl describe output and identify basic problems
  • Know the difference between ClusterIP, NodePort, LoadBalancer
  • Can do basic debugging: kubectl logs, kubectl exec
  • "You are here if you can": deploy an app, expose it, and check its logs without looking up syntax

Score 6-7

  • Understand the control loop: desired state vs actual state
  • Can debug CrashLoopBackOff, ImagePullBackOff, Pending pods
  • Understand RBAC: roles, bindings, service accounts
  • Know how to use resource requests/limits correctly
  • Can explain probes (liveness, readiness, startup) and when to use each
  • Comfortable with Helm values, templates, releases
  • Can perform rolling updates and rollbacks
  • "You are here if you can": diagnose why a pod is pending and fix it, explain the difference between liveness and readiness probes, and deploy with Helm

Score 8

  • Can perform node maintenance (cordon, drain, uncordon) safely
  • Understand PodDisruptionBudgets and when they matter
  • Can debug network policies, DNS issues, and service mesh problems
  • Know version skew rules for cluster upgrades
  • Understand etcd's role and can take snapshots
  • Can explain how the scheduler makes placement decisions
  • Know the difference between voluntary and involuntary disruptions
  • "You are here if you can": upgrade a cluster node without dropping traffic, explain why a network policy isn't working, and design a proper PDB strategy

Score 9

  • Deep understanding of kube-apiserver, etcd, scheduler, controller-manager internals
  • Can diagnose problems at the API server audit log level
  • Understand CRI, CNI, CSI plugin interfaces
  • Can debug kubelet issues, certificate expiration, control plane failures
  • Have dealt with split-brain, etcd quorum loss, or API server overload
  • "You are here if you can": recover a cluster from etcd backup, debug a kubelet that won't register, or explain the admission controller chain