Examkin

Question

Difficulty: HardDesigning Infrastructure for Business Requirements and Cost Optimization

A fashion retail enterprise operates an order fulfillment pipeline that renders personalized PDF invoices upon user checkout. The workload is completely stateless and experiences highly unpredictable traffic, ranging from zero requests during off-peak hours to sudden bursts of 300 requests per minute during promotional flash sales. The lead architect needs to select a compute architecture that minimizes overall infrastructure costs by scaling to zero during idle periods while eliminating container orchestration cluster management overhead. Which solution should the architect choose?

  1. Deploy the containerized application on Cloud Run with minimum instances set to 0.Answer
  2. B
    Provision a standard Google Kubernetes Engine (GKE) cluster with node auto-provisioning and deploy the application using the Horizontal Pod Autoscaler.
  3. C
    Deploy the application on a Compute Engine Managed Instance Group (MIG) backed by a 3-year Committed Use Discount (CUD) to cover baseline compute capacity.
  4. D
    Deploy the application on GKE Autopilot paired with a multi-region Cloud Spanner instance to handle traffic spikes.

Answer

Deploy the containerized application on Cloud Run with minimum instances set to 0.
Deploying on Cloud Run with minimum instances set to 0 fulfills all business requirements. Cloud Run automatically scales containers based on incoming web traffic and scales down to zero when no requests are being processed, eliminating idle compute costs. As a fully managed serverless product, it requires zero cluster management overhead.

Step-by-Step Solution

1
Analyze the workload characteristics and operational constraints.
The workload is stateless, highly variable with idle periods, and requires minimal administrative overhead.
Identifying that the workload spends time completely idle requires a compute platform capable of true scale-to-zero.
2
Evaluate Google Cloud compute options against cost and operational overhead criteria.
Cloud Run scales down to 0 instances when idle, incurring zero infrastructure charges during off-peak periods, and is fully managed without cluster administration.
Compute Engine MIGs and standard GKE clusters incur baseline node or cluster costs even when idle.
3
Select the optimal cost-effective architecture.
Cloud Run with minimum instances set to zero satisfies both the business cost optimization goal and operational simplicity requirement.
It handles rapid autoscaling during traffic bursts while charging exclusively per millisecond of active request processing.

Key Concept

Serverless Compute Right-Sizing and Scale-to-Zero Architecture
Rate this question

Topics

Designing Infrastructure for Business Requirements and Cost Optimization