Although it's an interesting calculation, you rarely have all your microservices communicating with eachother since you tend to group by areas. And even if that

Although it's an interesting calculation, you rarely have all your microservices communicating with eachother since you tend to group by areas. And even if that was the case, if good practices are in place (semantic versioning, good test cases, etc) I don't see why that should be a problem.