Examkin

Question

Difficulty: Very hardSelecting and Designing Compute and Application Platforms

A financial enterprise is architecting an online payment fraud risk evaluation service on Google Cloud. The workload consists of stateless containerized microservices that communicate via HTTP/2 gRPC. The service experiences extreme request volatility, surging to tens of thousands of requests per second during peak retail hours and dropping to near zero overnight. The leadership team requires a compute solution that minimizes operational management overhead, automatically scales down to zero instances during idle periods to optimize costs, and avoids managing underlying server infrastructure or cluster control planes. Which compute architecture should you recommend?

  1. Deploy the containerized service to Google Cloud Run, configuring autoscaling with minimum instances set to zero.Answer
  2. B
    Deploy the microservice on a Google Kubernetes Engine (GKE) Autopilot cluster using a Deployment with Horizontal Pod Autoscaler (HPA).
  3. C
    Deploy the microservice across a Compute Engine Managed Instance Group (MIG) behind an Application Load Balancer, configured with CPU utilization autoscaling.
  4. D
    Deploy the service on Compute Engine Virtual Machines backed by 3-year Committed Use Discounts (CUDs) sized for estimated peak traffic.

Answer

Deploying the microservice to Google Cloud Run with minimum instances set to zero satisfies all functional and operational constraints by providing fully managed container execution, native HTTP/2 gRPC support, automatic scale-to-zero cost savings, and zero cluster administration.
Deploying to Cloud Run is the optimal architectural choice because it provides a fully managed, serverless execution environment for containerized workloads. It natively supports HTTP/2 and gRPC, automatically handles high request bursts, scales down to zero when traffic stops, and eliminates all cluster maintenance, OS patching, and control plane management.

Step-by-Step Solution

1
Analyze workload requirements
Identified stateless containerized microservices requiring HTTP/2 gRPC protocol support and extreme autoscaling responsiveness.
The compute platform must support standard container images and gRPC traffic without requiring application re-architecture.
2
Evaluate operational and financial constraints
The solution must eliminate cluster management overhead and scale down to zero instances during idle periods.
Managing Kubernetes control planes or virtual machine OS patches increases operational overhead, while running static instances during low-traffic periods incurs unnecessary costs.
3
Select optimal GCP compute platform
Cloud Run meets all operational, architectural, and financial requirements as a fully managed serverless container runtime.
Cloud Run natively ingests gRPC traffic over HTTP/2, scales seamlessly from zero to thousands of instances, and abstracts away infrastructure management completely.

Key Concept

Selecting Serverless Container Runtimes (Cloud Run) vs. Kubernetes (GKE) for Stateless Microservices
Estimated Time:2m 0s
Rate this question

Topics

Selecting and Designing Compute and Application Platforms