Managed federation: Recommended development workflow


When using managed federation, what is the recommended workflow for local development?

Using the manual workflow, everything is fairly straight-forward since I need to be using rover to compose the subgraph, and this ties pretty easily into a typical graphql dev workflow(which usually involves quite a bit of code generation).

However, when using the managed workflow, every update to local subgraph requires a corresponding rover subgraph publish to update the local development gateway, right? If so, what is the recommended way to handle multiple developers? Is it just using variants?

Initially I had been thinking that using the managed workflow would also simplify the development setup since I would be able to eliminate rover supergraph compose, but instead it seems I will need an additional package.json task per microservice that publishes the subgraph on change.

I guess my question is: Is there some piece I’m missing here? Is there something that can simplify/streamline this process? Should I be setting up the manual workflow for local development only?