A smart utility company receives periodic HTTP telemetry payloads from 500,000 smart energy meters every 15 minutes. Traffic spikes predictably for 2 minutes during each 15-minute interval and remains virtually idle in between. The company needs to ingest these HTTP requests, normalize the telemetry payload, and store the structured records in a regional relational database for daily reporting. Which architecture minimizes baseline operational costs and management overhead while meeting the workload's performance requirements?
- Deploy a Cloud Run service to ingest and normalize the HTTP telemetry payloads, and write the processed records to Cloud SQL for PostgreSQL.Answer
- BProvision a multi-zonal Google Kubernetes Engine (GKE) Autopilot cluster running NGINX Ingress and stateless worker pods to process incoming HTTP requests into Cloud SQL for PostgreSQL.
- CDeploy a Cloud Run service to process incoming HTTP requests and persist the normalized telemetry records directly into a multi-region Cloud Spanner instance.
- DProvision a Compute Engine Managed Instance Group (MIG) sized for peak traffic capacity using 3-year Committed Use Discounts (CUDs) to ingest HTTP requests into Cloud SQL for PostgreSQL.
Answer
Deploying a Cloud Run service writing to Cloud SQL for PostgreSQL minimizes operational costs by leveraging scale-to-zero compute for bursty periodic HTTP telemetry payloads alongside a right-sized regional relational database.
Cloud Run is ideal for stateless HTTP workloads with intermittent or bursty traffic patterns because it scales down to zero instances during idle periods, preventing charges when no telemetry requests are arriving. Cloud SQL for PostgreSQL fulfills the regional relational reporting requirement cost-effectively without over-provisioning storage infrastructure.
Step-by-Step Solution
Key Concept
Serverless Compute & Database Selection for Periodic Workloads
Estimated Time:1m 30s