Discussion: GraphQL mutation granularity

Recently there have been some questions regarding mutations and how engineers should handle GraphQL mutations.

The broad questions:

• Should mutations be fine-grained ("update issue weight") or coarse-grained ("update issue")
• Should we set a standard for mutation granularity?

Our service-oriented architecture is such that most mutations that target a similar object call the same kind of service (in the example above, UpdateIssueService). So it's a matter of figuring out which we prefer.

I think it might be worth discussing the pros vs cons (on both frontend and backend) of both approaches.