I agree. Our architecture is partially built around logging every method called. Although it’s useful information, it still makes our product more difficult to program for.

Performance isn’t much of a problem though. Have different logging levels and only turn on a high level when you’re trying to nail down a particular STR.