Examkin

Question

Difficulty: HardSelecting and Designing Compute and Application Platforms

A healthcare enterprise is modernizing its telemetry processing service on Google Cloud. The application receives persistent WebSocket connections streaming real-time metric updates from medical IoT devices. The workload is stateless and packaged as a Docker container. During nighttime off-peak hours, device traffic drops to zero, and the platform must automatically scale down to zero active instances to eliminate infrastructure costs. During daytime operational hours, the service experiences sudden traffic spikes and must eliminate cold-start latency for incoming connections while minimizing operational overhead by using a fully managed serverless infrastructure. Which compute architecture should the Lead Cloud Architect recommend?

  1. Deploy the containerized service on Cloud Run, configuring request concurrency to accommodate multiple WebSocket connections per instance, setting minimum instances during operational hours, and allowing autoscaling to zero instances during off-peak hours.Answer
  2. B
    Deploy the containerized service on a Google Kubernetes Engine (GKE) Standard cluster configured with a Horizontal Pod Autoscaler (HPA) and Cluster Autoscaler to scale pods and nodes down to zero during idle periods.
  3. C
    Deploy the containerized service on a Compute Engine Managed Instance Group (MIG) behind an Application Load Balancer using an autoscaling policy based strictly on average CPU utilization set to 80%.
  4. D
    Deploy the containerized service on App Engine Flexible Environment while purchasing 3-year Committed Use Discounts (CUDs) for the underlying Compute Engine instances to cover baseline traffic.

Answer

Deploying the containerized service on Cloud Run with high request concurrency, minimum instances configured during operational hours, and scale-to-zero enabled off-peak meets all technical, operational, and cost requirements.
Cloud Run is the optimal Google Cloud compute platform for stateless containerized applications that require zero operational infrastructure management. It natively supports WebSockets and allows multiple concurrent requests per instance. By configuring minimum instances during peak operational hours, cold starts are eliminated. During off-peak hours, setting minimum instances to zero allows Cloud Run to scale down completely, incurring zero compute cost.

Step-by-Step Solution

1
Analyze operational overhead requirements
Fully managed serverless platforms (such as Cloud Run) eliminate cluster management and node provisioning overhead compared to GKE or Compute Engine MIGs.
The enterprise requirement specifies zero VM or cluster management overhead.
2
Evaluate workload scaling and protocol requirements
Cloud Run natively supports stateless containerized workloads, WebSocket protocols, multi-concurrency per instance, and autoscaling down to zero instances.
Persistent WebSocket connections demand high concurrency per instance and seamless scale-to-zero support during off-peak hours.
3
Address latency and cold-start constraints during peak hours
Setting Cloud Run minimum instances during daytime operational hours ensures pre-warmed instances handle incoming traffic without cold-start spikes.
Minimum instances maintain baseline readiness during operational hours while permitting complete scale-to-zero overnight.

Key Concept

Selecting Cloud Run for stateless containerized workloads requiring serverless operations, WebSocket protocol support, custom concurrency, and cold-start mitigation via minimum instances.
Rate this question

Topics

Selecting and Designing Compute and Application Platforms