Reduce latency for Asian users of a stateless web API currently hosted in a single us-central1 GKE cluster.

Creating a second GKE cluster in asia-southeast1 and using kubemci to provision a global HTTP(S) Load Balancer across clusters is the most effective solution. This approach brings the application closer to Asian users, reducing network latency. kubemci (Kubernetes Multi-Cluster Ingress) simplifies the deployment and management of a single global HTTP(S) Load Balancer that routes traffic to services across multiple GKE clusters, providing a unified entry point and optimal routing. Using a global HTTP(S) Load Balancer with Cloud CDN enabled is a good option for static content, but for a stateless web API, the backend proximity is more critical than caching. Increasing CPU/memory only scales the existing cluster, not addressing geographical latency. Exposing services with Service type LoadBalancer and adding IPs to Cloud DNS would require manual DNS-based load balancing, which is less efficient and robust than a global HTTP(S) Load Balancer.