GCP ACE Practice / Study guide / Topics / Google Kubernetes Engine (GKE)
What actually gets tested on the Google Cloud Associate Cloud Engineer exam, the traps to know, and 6 free sample questions with full explanations.
Creating a cluster is gcloud container clusters create; the flag that matters most conceptually is --region versus the zonal default — --region makes a regional (multi-zone) cluster where --num-nodes applies per zone, which is a common source of "why do I have more nodes than I asked for" confusion in scenario questions. Once the cluster exists, gcloud container clusters get-credentials writes its endpoint and auth config into your local kubeconfig so kubectl can talk to it — logging in again doesn't add a cluster context, and it's a frequent wrong-answer distractor.
Autopilot vs Standard is the exam's other core GKE decision: Autopilot is fully managed Kubernetes where Google operates the nodes and you pay per pod, trading some control for zero node management; Standard leaves node pool sizing, upgrades, and maintenance to you in exchange for more control. "Minimize operational overhead" in a GKE question usually points to Autopilot the same way it points to Cloud Run in a broader compute-choice question.
On the workload side, know what each object is for without needing deep YAML mastery: Deployments manage ReplicaSets to keep a desired replica count running and orchestrate rolling updates/rollbacks; Services give pods stable networking, where type: LoadBalancer specifically provisions a Google Cloud passthrough load balancer with an external IP (Layer 4) — for HTTP(S) routing you'd reach for an Ingress instead, and the exam tests that L4-vs-L7 distinction directly.
These 6 questions are drawn from our full bank of 658 to show the style and depth you'll get in the app. Each comes with the full explanation you'd see after answering in a real practice session.
Your team needs a basic GKE cluster named apps with 3 nodes per zone across a region. Which command creates it?
A teammate asks who is responsible for upgrading the Kubernetes control plane of your GKE Standard cluster. How do control-plane upgrades happen?
Right after creating your GKE cluster, kubectl on your laptop complains it has no configured context. What do you run?
A node in your GKE cluster has been failing its health checks for several minutes. What does node auto-repair do about it?
Your stateless web pod must always run 4 replicas and support rolling updates of its image. Which Kubernetes object manages this?
A pod in your cluster is stuck in CrashLoopBackOff and you need the last 100 log lines from the crashed container instance. Which command?
The app has dozens more questions on this and every other topic, with instant scoring, spaced repetition of what you get wrong, and progress tracking across sessions.
Start practicing →