Question

Difficulty: MediumPlanning and Assessing Cloud Compute Solutions

A company is planning to deploy a stateless containerized web service packaged as a custom Docker image. The workload experiences unpredictable HTTP traffic spikes, requires automatic scaling down to zero instances during idle periods to minimize operational costs, and must be deployed with zero virtual machine or Kubernetes cluster management overhead. Which Google Cloud compute platform best meets these requirements?

  1. Cloud RunAnswer
  2. B
    Cloud Functions
  3. C
    Google Kubernetes Engine (GKE) Standard
  4. D
    Compute Engine Spot VMs in a Managed Instance Group

Answer

Cloud Run is the optimal compute platform for containerized stateless web applications that require automatic scaling to zero with zero cluster management overhead.
Cloud Run is a fully managed serverless platform that deploys stateless container images directly. It automatically scales instances up to match incoming HTTP traffic and scales down to zero when no requests are being processed, completely eliminating cluster administration and idle compute costs.

Step-by-Step Solution

1
Analyze workload characteristics
The application is stateless, packaged as a Docker container, and processes HTTP requests with fluctuating demand.
Containerization and HTTP request pattern make the application eligible for container-based platforms.
2
Evaluate operational and cost requirements
The solution requires scaling to zero during idle times and zero infrastructure/cluster management overhead.
This rules out IaaS (Compute Engine) and managed Kubernetes clusters (GKE Standard) that require node management.
3
Select the appropriate GCP serverless platform
Cloud Run runs arbitrary Docker containers serverlessly while scaling to zero and eliminating node maintenance.
Cloud Run directly satisfies container portability and zero-management requirements better than function-based serverless runtimes.

Key Concept

Selecting Cloud Compute Services Based on Workload Characteristics and Operational Overhead
Estimated Time:1m 30s
Rate this question