A global digital publishing enterprise is re-architecting its core media platform on Google Cloud. The system consists of a stateless HTTP API backend service and a transactional relational database holding article metadata and subscription state. The technical requirements mandate 99.99% multi-regional uptime, a strict zero Recovery Point Objective () for transactional database updates across regions, and automatic traffic routing away from an impaired region. Which TWO architectural components should you select to satisfy these requirements?
- Deploy the stateless API service to Cloud Run across multiple GCP regions and front them with a Global External Application Load Balancer.Answer
- Use a Cloud Spanner database with a multi-region instance configuration for article metadata and subscription state.Answer
- CDeploy a regional Google Kubernetes Engine (GKE) cluster in a single region and rely on manual cluster node re-provisioning for disaster recovery.
- DProvision a Primary Cloud SQL instance with cross-region asynchronous read replicas to achieve zero multi-region RPO.
- EConfigure the HTTP load balancer health check path to execute a synchronous SELECT query against the relational database on every ping.
Answer
The platform should be architected using Cloud Run deployed across multiple GCP regions behind a Global External Application Load Balancer alongside a multi-region Cloud Spanner database instance.
Combining multi-region Cloud Run services under a Global External Application Load Balancer ensures traffic is automatically directed to healthy regional endpoints. Using Cloud Spanner with a multi-region configuration provides multi-region synchronous transactional replication, satisfying the requirement for and high availability.
Step-by-Step Solution
Key Concept
Designing Multi-Region High Availability with Zero RPO