Question

Difficulty: EasyPlanning Google Kubernetes Engine (GKE) Cluster Architectures

A cloud administrator is planning a new Google Kubernetes Engine (GKE) cluster deployment for a suite of web microservices and wants to reduce node management overhead while optimizing resource utilization. The administrator considers provisioning the cluster in GKE Autopilot mode. Which TWO statements accurately describe the operational characteristics and billing model of GKE Autopilot? (Select TWO options.)

  1. Google manages node provisioning, scaling, auto-repair, and node security patching automatically.Answer
  2. Billing is based directly on the CPU, memory, and ephemeral storage resources requested by running Pods rather than provisioned Compute Engine VM capacity.Answer
  3. C
    Cluster administrators must manually install and configure the Cluster Autoscaler tool to add worker nodes when workloads expand.
  4. D
    Custom Linux kernel parameters (sysctls) and custom node operating system packages can be freely modified by cluster operators on underlying nodes.
  5. E
    Spot workloads deployed on Autopilot are guaranteed against preemption for stateful databases.

Answer

Google fully manages node provisioning and security, and billing is calculated based on Pod resource requests rather than total node capacity.
GKE Autopilot abstracts away node management by delegating node provisioning, repair, and security patching to Google. Furthermore, its cost model bills based on the Pod's resource requests rather than paying for unutilized node headroom.

Step-by-Step Solution

1
Analyze GKE Autopilot management responsibilities
Autopilot provides a hands-off experience where Google manages all node infrastructure, auto-scaling, operating system updates, and node security hardening.
This directly fulfills the requirement to minimize operational and node management overhead.
2
Analyze the GKE Autopilot pricing model
GKE Autopilot charges users based on the CPU, memory, and storage resource requests specified in the Pod manifests rather than charging for provisioned Compute Engine VM instance sizes.
This guarantees that users pay only for the resources requested by their workloads, avoiding charges for unallocated node capacity.

Key Concept

GKE Autopilot Cluster Architecture and Resource Billing Model
Rate this question