A smart energy grid company ingests continuous HTTP telemetry streams from millions of regional meters and executes scheduled nightly batch processing for billing calculations. Raw telemetry payloads must be retained for years for regulatory compliance but are accessed less than once per year after days. The company's database stores regional device state metadata requiring standard relational SQL querying and high availability, but does not require global multi-region write synchronization. The chief architecture officer has set strict business mandates to minimize baseline compute management overhead, eliminate unnecessary database licensing costs, and reduce long-term cold storage expenditure. Which TWO architectural decisions should you recommend? (Select TWO.)
- Deploy the stateless HTTP telemetry ingestion microservices on Cloud Run, and set Cloud Storage Lifecycle Management rules to transition raw telemetry objects to Coldline Storage after 30 days and Archive Storage after 365 days.Answer
- Provision regional Cloud SQL for PostgreSQL with High Availability for device metadata, and execute nightly batch processing jobs using Spot VMs or Cloud Run Jobs.Answer
- CDeploy the stateless HTTP ingestion services on a Dedicated GKE Autopilot cluster across three regions to guarantee container resource reservation for all ingress traffic.
- DProvision a multi-region Cloud Spanner instance to store device metadata, ensuring continuous cross-continent synchronous write scalability.
- EPurchase 3-year Committed Use Discounts (CUDs) covering peak compute capacity for all nightly batch processing virtual machines on standard Compute Engine VMs.