python logger的后代-好处

Question

I'm having trouble to understand the benefits of creating descendants loggers out of my root logger, in each module. For instance:

In project called "foo", module name called "bar":

import logging
logger = logging.getLogger(__name__)
# other pieces of code, that all of them use log...

And same project, but in module "baz"

import logging
logger = logging.getLogger(__name__)
# other pieces of code, that all of them use log...

instead of just, creating one logger object, and make other modules import and use it. For instance in same project:

from foo.logs import logger
# other pieces of code, that all of them use log...

and foo/logs.py would contain something like:

import logging
logger = logging.getLoger('foo')
# some init logic of the logger...

I can get all data I need from the created LogRecord object: filename, funcName, pathname, name, etc...

Why repeat same procedure of creating a descendant logger (calling getLogger) when I use one? it may also have a minor performance footprint no?

Answer 1

The reason for creating a separate logger for each file is so that you can later customize what is and what isn't logged without changing the source code.

Logging is usually configured via an external file, which can be modified after the installation is installed, even by users who are not software developers.

For the same reason, there are different logging levels which can be used with each logger, in order to allow more or less logs to be produced. Then at some point, when eg exhaustive log data from module a is needed for debugging purposes, the configuration file can be modified to show DEBUG log level for module a without cluttering up the log file with debug info from every module in the application.