An enterprise SaaS company is planning the architecture for a new containerized web application on Google Kubernetes Engine (GKE). The application consists of stateless microservices that scale dynamically based on HTTP traffic. The infrastructure team wants to eliminate manual node provisioning, OS patching, and capacity planning while ensuring Google manages the cluster infrastructure health and node lifecycle entirely. Additionally, organizational policy forbids running pods with privileged host access. Which GKE mode should the cloud engineer choose to meet these requirements?
- Provision a GKE Autopilot cluster, allowing Google Cloud to fully manage the underlying node infrastructure and enforce built-in security baselines.Answer
- BProvision a GKE Standard cluster with node auto-provisioning enabled, because Autopilot requires operators to manually install node operating system security patches.
- CProvision a GKE Standard cluster using Spot VM node pools exclusively for the control plane nodes to eliminate control plane management costs.
- DProvision a GKE Standard cluster with a single zonal node pool and disable the Cluster Autoscaler to eliminate cluster management overhead.
Answer
Provision a GKE Autopilot cluster, allowing Google Cloud to fully manage the underlying node infrastructure and enforce built-in security baselines.
GKE Autopilot is designed to provide a fully managed environment where Google handles the cluster infrastructure, including node provisioning, auto-scaling, auto-repair, and OS patching. It also enforces security best practices by blocking privileged pod execution by default, making it the ideal choice for stateless microservices with zero node administration overhead.
Step-by-Step Solution
Key Concept
GKE Autopilot vs. Standard Cluster Architecture Planning