A cloud administrator is planning a new Google Kubernetes Engine (GKE) cluster to host stateless web microservices. The primary objective is to minimize node management overhead so that Google fully manages the underlying worker node infrastructure and OS patching, while charging only for the resource requests of deployed Pods. Which TWO architectural choices or operational features should the administrator select to fulfill these requirements?
- Deploy the cluster using GKE Autopilot mode.Answer
- Define explicit CPU, memory, and storage resource requests for each Pod specification.Answer
- CDeploy the cluster using GKE Standard mode with custom node pools.
- DUse Spot VMs for all node pools hosting persistent stateful database workloads.
- EConfigure Horizontal Pod Autoscaler (HPA) to automatically provision additional Compute Engine VM nodes.
Answer
The administrator should deploy the cluster using GKE Autopilot mode and specify explicit CPU, memory, and storage resource requests in the Pod manifests.
GKE Autopilot is Google Cloud's fully managed cluster mode where Google handles all node management, security patching, and infrastructure scaling. In Autopilot, billing is based directly on the CPU, memory, and storage requested in individual Pod manifests, eliminating payment for unused node capacity.
Step-by-Step Solution
Key Concept
GKE Autopilot cluster architecture and Pod-based resource planning