I once worked on a program that generated 2 gigs of logs a day, and 99% of them were useless, and only a couple of people in the organization could even debug these log files..

Don't get me wrong - even crappy logging is invaluable for debugging, you can't get away without it on a production project, but you need to be very careful to only add logging that might in fact be useful.

A second quibble, I'm sure you're sick of me now :-D, is that Python's logging is antiquated. It requires you to use %-format strings, which are in every way inferior to format strings, and make common things like "teeing a log so it goes to the console and to a file" tricky to get working.

But much more important than that is that text logs were obsolete twenty years ago. How are you going to process them? Are you going to use grep to find things? How are you going to do analytics on those logs?

Now the cool kids all uses structured logs, where each entry is in some format like JSON so an automatic program can parse it.