Considering Errors in Dependent Services for Service Definitions
Regarding services. Should the definition of a service account for potential errors of dependent services ? My intuition is that they should not, as you define your service with just the ideal and specific interface and leave the rest to the implementation. However, the implementation will probably depend on another service with it's own set of possible errors (let's imagine a User service and a DB service). How are those errors propagated in the type? Any part of the code using the user service will be oblivious of any DB error, so I assume those errors will be pushed to the boundaries? But how?
