There's nothing wrong with the architecture you describe. In fact, I've even migrated/merged microservices into a monolith several times. What you describe abou

There's nothing wrong with the architecture you describe. In fact, I've even migrated/merged microservices into a monolith several times. What you describe about restarting, etc. is accurate, and you need to solve for that even in other architectures. Conway's law is more important: "I need to deploy my team's change without waiting for your team's job to finish".