The very first thing I do when starting a new project is putting the logging infrastructure in place. In production you don't want to spam stderr with each and every event. But if some operation fails the best way to understand why, is having a precise log of what chain of operations lead to this error.