Question

Difficulty: MediumSelecting and Designing Compute and Application Platforms

An enterprise energy platform is deploying a new stateless microservice to process inbound HTTP REST telemetry payloads from solar inverters. The workload experiences zero traffic at night and sharp traffic spikes during daylight hours. The architectural team requires auto-scaling to zero to minimize costs and has mandated minimizing operational management overhead. 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) Standard cluster.
  3. C
    Deploy the service on a Compute Engine Managed Instance Group (MIG) configured with target CPU utilization auto-scaling.
  4. D
    Deploy the service to a Compute Engine instance pool using 3-year Committed Use Discounts (CUDs).

Answer

Deploy the containerized service to Cloud Run.
Cloud Run is the optimal Google Cloud compute choice for stateless HTTP web microservices. It abstracts server management, automatically scales compute instances based on incoming request traffic (including scaling to zero when idle), and eliminates infrastructure operational overhead.

Step-by-Step Solution

1
Analyze workload requirements
The service is stateless, communicates via HTTP REST, requires auto-scaling to zero during non-daylight hours, and must minimize operational overhead.
Matching workload operational characteristics directly determines the optimal GCP compute service.
2
Evaluate candidate compute platforms against constraints
Cloud Run supports containerized HTTP workloads, scales automatically down to zero instances, eliminates node management, and bills strictly per execution resource unit.
Serverless container execution fulfills both the cost optimization and zero-operational-overhead mandates.

Key Concept

Selecting serverless container infrastructure (Cloud Run) over GKE or VM instance groups for stateless HTTP workloads to minimize operational overhead and scale to zero.
Rate this question