Yes, you are correct, there is not an answer and there is not much data on pros and cons. The main point in the article is precisely this: do not get stuck in a concept that is not yet mature enough.

Like I said in the text, we can have a point in the project that consumes a lot of processing, but with little access to data. I have a real case like that. So it makes sense to use a microservice to be able to scale just that service.

About make, I have another post explaining the DevOps approach.