Logging

Links

From Become a logging expert in 30 minutes:

The logger is the point of entry for all logged messages.
logger is namespace specific.
It decides whether it is going to accept or reject the message.
The root logger is the top level piece. propagate = True. propagate says to take this setting and apply it to every single key in the namespace underneath e.g. if we set propagate = True for foo, then the setting will apply to foo.bar.baz.
If you get double logging, then check propagate = True.
incremental = True, adds to the pre-existing configuration. This has dangerous behaviour i.e. you can’t be sure of the pre-existing configuration.
Handlers deliver messages to their destination.
Formatters are destination specific.
Filters are applied to loggers e.g. if you want to log output from one IP address.

Tip

YAML config looks very clean. Check the video for an example.

Setup

Create a module to manage the configuration:

import logging

LOGGING_FORMAT = '%(asctime)s %(levelname)-5s %(module)s %(funcName)s %(message)s'

logging.basicConfig(
    level = logging.DEBUG,
    format = LOGGING_FORMAT,
    datefmt = '%H:%M:%S',
    filename = 'logger.log',
    filemode = 'w'
)

Note:

See Formatter Objects.
See LOGGING_FORMAT
Remove the filemode if you want to append to the file.
If (for example) you name this module, logging_config.py, your main module (or unit test), will need to import logging_config to activate the configuration.
python version 2.4, didn’t seem to recognise the funcName parameter.

In the other modules of your application, import the configuration module and start logging:

import logging

logger = logging.getLogger(__name__)

logger.debug('some log text')

See Advanced Logging Tutorial for details about using getLogger.