Logger wrapper allowing one to stream logs to both stdout and Azure Application Insights.

It uses python's standard library logging capability to expose custom handlers and formatters for communication with Azure Application Insights.

In order to use just before your first usage of the logger (where you call logging.getLogger()) make a call the following call:

from logger.config import config_loggers

config_loggers(
    app_insights_key='<YOU.KEY>',
    app_insights_app_id='<APP.ID>',
    app_version='<APP.VERSION>',
    app_name='<APP.NAME>',
    correlation_id='<CORRELATION.ID>',
    execution_env_name='DataBricks'):

logger = logging.getLogger()

logger.info('SOMETHING_HAPPENED', {'start': 123})

Also if the configuration delivered with the config_loggers is not fulfilling all criteria one can use is a merely an inspiration and provide ones own customization.

The underlying formatter and handler is expecting the logs to be invoked in a particular format:

logger('SOME_EVENT_NAME', {'here': 'extra data'})

for example:

logger(
    'ARES_JOB_SUCCEEDED',
    {'count': 11, 'datetime': '2018-11-12T12:32:31'})

Where all numerical values will end up in the measurements part of Application Insights and all non-numerical will end up in the properties part.

Also the properties will be enriched by some meta information allowing one to understand the invocation context.