I saw a request a while ago asking for advice on some kind diagnostic logging (I can’t recall the details other than the person who asked). Worded another way the question is: how do apply a cross-cutting concern?
When we:
- design our solutions around the ideas of queues, messages, events and streams & services (aka queries & commands)
- write code with semantic units of abstraction utilising generic variance
We no longer need to ask how, we simply decide where.