It's not about replication, it's about compartmentalizing code changes / bugs / resource utilization. If you deploy a bug that causes a service crash loop or re

It's not about replication, it's about compartmentalizing code changes / bugs / resource utilization. If you deploy a bug that causes a service crash loop or resource exhaustion isolating that to a small service reduces impact to other services. And if that service isn't core to app then the app can still function.