Examkin

Question

Difficulty: MediumPlanning Google Kubernetes Engine (GKE) Cluster Architectures

A financial services company is architecting a new Google Kubernetes Engine (GKE) cluster to host a stateless REST API. The infrastructure management team requires a cluster architecture that completely eliminates worker node management overhead—including node provisioning, scaling, and OS security patching—while billing strictly for the CPU, memory, and storage resources requested by running Pods rather than unutilized compute node capacity. Which GKE cluster architecture best fulfills these requirements?

  1. Deploy a GKE Autopilot cluster.Answer
  2. B
    Deploy a GKE Standard cluster with Cluster Autoscaler enabled across all node pools.
  3. C
    Deploy a GKE Standard cluster utilizing Spot VMs for the primary node pool.
  4. D
    Deploy a GKE Standard cluster and configure Horizontal Pod Autoscaler (HPA) to manage node capacity.

Answer

Deploying a GKE Autopilot cluster meets all requirements by automating node infrastructure management and billing based on Pod resource requests.
GKE Autopilot fully manages cluster infrastructure including node provisioning, auto-scaling, auto-upgrading, and node security hardening. Furthermore, billing in Autopilot mode is calculated based on the CPU, memory, and ephemeral storage requested by deployed Pods rather than provisioned Compute Engine VM capacity.

Step-by-Step Solution

1
Analyze the operational requirements.
Identified two main requirements: zero node management overhead (provisioning, patching, scaling) and Pod-based request billing.
Choosing between GKE Standard and GKE Autopilot depends on operational responsibility boundaries and billing models.
2
Compare GKE Autopilot and GKE Standard operational modes.
GKE Autopilot automates node management and bills strictly per Pod resource request. GKE Standard requires node pool management and charges per compute VM node capacity.
GKE Autopilot abstracts the underlying Compute Engine instances completely from the administrator.
3
Select the matching cluster mode.
GKE Autopilot is the only option that eliminates node management overhead while matching the requested Pod resource billing model.
Standard GKE clusters, even with autoscaling features like Cluster Autoscaler or HPA, retain node-level configuration and VM capacity billing.

Key Concept

GKE Autopilot vs. Standard Operational Boundaries
Estimated Time:1m 30s
Rate this question

Topics

Planning Google Kubernetes Engine (GKE) Cluster Architectures

More questions from this topic