You broken a monolith into containerized REST microservices and want to run them on Cloud Run with high availability and low latency for customers. What should you do?

To achieve high availability and low latency globally, deploying Cloud Run services to multiple regions is essential. Cloud Run services are regional resources, so deploying to multiple regions inherently provides resilience against regional outages. Serverless Network Endpoint Groups (NEGs) are specifically designed to integrate serverless backends like Cloud Run with Google Cloud's global HTTP(S) Load Balancer. The global HTTP(S) Load Balancer then distributes traffic to the nearest healthy region, ensuring low latency and high availability. Deploying to multiple availability zones is insufficient for global low latency and high availability, as zones are within a single region. Cloud Endpoints are for API management, not direct load balancing. Latency-based DNS routing can introduce caching issues and doesn't offer the same traffic management capabilities as a global HTTP(S) Load Balancer. A TCP/IP global load balancer is for TCP traffic, not HTTP(S) services like REST microservices.