Examkin

Question

Difficulty: MediumSelecting and Designing Compute and Application Platforms

A global retail enterprise is building a lightweight, stateless microservice to handle peak spikes of HTTP inventory availability queries during high-volume promotional events. Traffic to this microservice exhibits extreme variability, surging from zero requests during off-peak hours to tens of thousands of HTTP requests per second during flash sales. The engineering team requires a containerized execution environment that automatically scales down to zero instances to minimize costs, requires zero underlying server or cluster management overhead, and charges strictly per millisecond of request processing time. Which Google Cloud compute platform should the Cloud Architect select?

  1. Deploy the containerized service to Cloud Run.Answer
  2. B
    Deploy the containerized service to a Google Kubernetes Engine (GKE) Autopilot cluster.
  3. C
    Deploy the service on a Compute Engine Managed Instance Group (MIG) configured with CPU-based autoscaling.
  4. D
    Deploy the service using App Engine Flexible Environment.

Answer

Deploying the microservice to Cloud Run satisfies all requirements for stateless execution, automatic scaling to zero, zero cluster management overhead, and request-based billing.
Cloud Run is designed specifically for stateless containerized workloads. It scales instances automatically up and down (including scaling down to zero when idle), eliminates infrastructure and cluster administration overhead, and charges only for compute resources consumed while processing requests.

Step-by-Step Solution

1
Analyze workload characteristics and requirements
Workload is a stateless containerized HTTP microservice with bursty, unpredictable traffic scaling from zero to thousands of requests per second.
Determining scaling dynamics and protocol needs narrows down compute options.
2
Evaluate Google Cloud serverless and container execution platforms
Cloud Run native container deployment provides serverless scaling to zero, zero node management, and per-millisecond request billing.
Matching technical requirements to GCP platform capabilities identifies the most operationally efficient choice.
3
Eliminate platforms with cluster overhead or baseline infrastructure costs
GKE, Compute Engine MIGs, and App Engine Flex require continuous infrastructure maintenance or minimum baseline costs during idle traffic periods.
Eliminating options that introduce management overhead or idle cost fulfills cost-optimization and operational directives.

Key Concept

Compute Platform Selection for Stateless Bursty Microservices
Rate this question

Topics

Selecting and Designing Compute and Application Platforms