A healthcare software provider processes nightly batch export files (~50 GB daily) containing electronic health records. The processing pipeline executes heavy data transformations for 45 minutes each midnight. The transformed output must be stored in a relational database where internal business analysts run daily analytical reports during standard business hours. The company wants to minimize infrastructure costs while meeting all operational requirements. Which TWO architectural decisions should you recommend?
- Execute the nightly transformation tasks using serverless Cloud Run jobs that scale down completely upon completion.Answer
- Store transformed data in Cloud SQL for PostgreSQL to serve internal business reporting queries.Answer
- CDeploy a dedicated 24/7 Google Kubernetes Engine (GKE) Autopilot cluster to run the nightly transformation workload.
- DProvision a multi-region Cloud Spanner instance to store the transformed data for reporting access.
Answer
The recommended approach combines serverless Cloud Run jobs for on-demand 45-minute nightly processing and Cloud SQL for PostgreSQL to host internal reporting data cost-effectively.
Selecting serverless Cloud Run jobs ensures compute resources consume zero budget during the 23+ idle hours per day. Pairing this with Cloud SQL for PostgreSQL delivers required relational database features for business analytics without incurring high multi-region global database baseline fees.
Step-by-Step Solution
Key Concept
Optimizing infrastructure costs by matching workload duty cycles to serverless compute paradigms and right-sizing relational storage services.