Python logging configurations done The Right Way

TL;DR:

logging.basicConfig() is very often not enough, leading to DRY violations. Use service-logging to log to syslog or the Windows event log when running in the background, and log to stderr when running in the foreground, both with well formatted messages.

Using logging correctly is often both usage and OS specific. For a usage specific example, when developing or troubleshooting a daemon and/or service it is often useful to use one logging configuration when running in the foreground in a shell and another when running in the background. For an OS specific example, when running in the background a daemon and/or service should use the appropriate logging service for that system so that logging metadata is correct (severity, facility, etc.), but the right way to do that is OS specific. Furthermore, deciding between these configurations requires logic that cannot be done through the simple configuration mechanisms provided by Python’s logging.config.

Given these realities, logging is often not done “The Right Way”, and/or is difficult to adjust for debugging/testing vs background. service-logging aims to address this by providing the logic to choose between correct logging configurations for different usages and OSes. It also provides several ways to use these configurations, making the depth of commitment opt-in.

$ sudo pip3 install service-logging

>>> import servicelogging
>>> servicelogging.basicConfig()

$ python -m servicelogging foreground_demo.py
$ service-logging foreground_demo.py

$ service-logging --help
usage: service-logging [-h] [--level LEVEL] script

python logging configurations done The Right Way, top-level package.

positional arguments:
  script         The Python script to run after configuring logging

optional arguments:
  -h, --help     show this help message and exit
  --level LEVEL  The level of messages to log at or above (default: 20)

>>> import servicelogging
>>> handler = servicelogging.choose_handler()