I have a Python Flask application, the entry file configures a logger on the app, like so:

app = Flask(__name__)
handler = logging.StreamHandler(sys.stdout)

I then do a bunch of logging using

app.logger.debug("Log Message")

which works fine. However, I have a few API functions like:

@app.route('/api/my-stuff', methods=['GET'])
def get_my_stuff():
    db_manager = get_manager()
    query = create_query(request.args)

    service = Service(db_manager, query)
    app.logger.debug("Req: {}".format(request.url))

What I would like to know is how can I do logging within that Service module/python class. Do I have to pass the app to it? That seems like a bad practice, but I don’t know how to get a handle to the app.logger from outside of the main Flask file…

Even though this is a possible duplicate I want to write out a tiny bit of python logging knowledge.

DON’T pass loggers around. You can always access any given logger by logging.getLogger(<log name as string>). By default it looks like* flask uses the name you provide to the Flask class.

So if your main module is called ‘my_tool’, you would want to do logger = logging.getLogger('my_tool')in the Service module.

To add onto that, I like to be explicit about naming my loggers and packages, so I would do Flask('my_tool')** and in other modules, have sub level loggers like. logger = logging.getLogger('my_tool.services') that all use the same root logger (and handlers).

* No experience, based off other answer.

** Again, don’t use flask, dk if that is good practice

Edit: Super simple stupid example

Main Flask app

import sys
import logging

import flask

from module2 import hi

app = flask.Flask('tester')

handler = logging.StreamHandler(sys.stdout)
    '%(asctime)s - %(name)s - %(levelname)s - %(message)s'))

def index():
    return "hi"

if __name__ == '__main__':


import logging

log = logging.getLogger('tester.sub')

def hi():
    log.warning('warning test')

Outputs - - [04/Oct/2016 20:08:29] "GET /index HTTP/1.1" 200 -
2016-10-04 20:08:29,098 - tester - DEBUG - TESTING!
2016-10-04 20:08:29,098 - tester.sub - WARNING - warning test

Edit 2: Messing with subloggers

Totally unneeded, just for general knowledge.

By defining a child logger, done by adding a .something after the root logger name in logging.getLogger('root.something') it gives you basiclly a different namespace to work with.

I personally like using it to group functionality in logging. So have some .tool or .db to know what type of code is logging. But it also allows so that those child loggers can have their own handlers. So if you only want some of your code to print to stderr, or to a log you can do so. Here is an example with a modified module2.


import logging
import sys

log = logging.getLogger('tester.sub')
handler = logging.StreamHandler(sys.stderr)
handler.setFormatter(logging.Formatter('%(name)s - %(levelname)s - %(message)s'))

def hi():

Output - - [04/Oct/2016 20:23:18] "GET /index HTTP/1.1" 200 -
2016-10-04 20:23:18,354 - tester - DEBUG - TESTING!
tester.sub - WARNING - test
2016-10-04 20:23:18,354 - tester.sub - WARNING - test