A financial analytics company executes scheduled daily risk-assessment batch computations for corporate clients. The batch job runs for approximately two hours every night, requiring dynamic scaling based on client volume. The application stores financial ledger transaction records in a standard relational database with regional availability requirements. Currently, the infrastructure relies on continuously running Virtual Machines and over-provisioned database instances, incurring high idle costs. You need to redesign the solution architecture to minimize compute and database costs while meeting technical requirements. Which two architectural decisions should you recommend? (Select TWO.)
- Execute the daily batch calculations using Cloud Run Jobs so compute resources automatically scale to zero when jobs complete.Answer
- BProvision a dedicated Google Kubernetes Engine (GKE) cluster with cluster autoscaling to execute the nightly two-hour batch workload.
- Deploy Cloud SQL with High Availability (HA) enabled for the relational database requirement.Answer
- DMigrate the ledger database to a multi-region Cloud Spanner instance to ensure regional fault tolerance.
Answer
The optimal architecture combines Cloud Run Jobs for containerized batch execution scaling to zero, and Cloud SQL with High Availability (HA) for regional relational storage.
Executing intermittent batch tasks via Cloud Run Jobs ensures compute resources are billed only for execution time and scale to zero when idle. Utilizing Cloud SQL with High Availability fulfills regional relational database durability requirements at a significantly lower cost than multi-region horizontal database offerings.
Step-by-Step Solution
Key Concept
Selecting serverless compute for intermittent batch workloads and right-sizing relational storage databases to balance cost and availability.