python-logger
Supplements a default formatter configuration for machine-readable JSON logging and applies it.
This module should be included as a sub-module in python projects with logging.
Please avoid copy-pasting as no updates can be supplied that way.
Usage
- Install the module via the
requirements.txt
or per CLI: pip install python-json-log-formatter
- Please include the
PythonLogger
module, using a from module import class
statement.
- After the imports, please declare a version constant.
This constant should use semantic versioning, with a prepended date (YYYY/MM/DD)
.
Other examples, without the date, can be found here
- Before any argument parsing or any other logging, the logger should be initialized
- In EVERY file, the logger needs to be imported using
LOGGER = logging.getLogger(__name__)
from python_json_log_formatter import PythonLogger
import logging
VERSION = "1.0.0 (2022/11/24)"
PythonLogger.setup_logger(VERSION)
LOGGER = logging.getLogger(__name__)
Optionally, it is also possible to change the log level:
from python_json_log_formatter import PythonLogger
import logging
VERSION = "1.0.0 (2022/11/24)"
PythonLogger.setup_logger(VERSION, logging.DEBUG)
LOGGER = logging.getLogger(__name__)
Features
Sets the logging format for the root logger and thus for every child logger.
In EVERY file where logging happens, please use LOGGER = logging.getLogger(__name__)
to get an individual logger.
Allows printing exception information on any logging level, not only on the EXCEPTION
level but also for INFO
or DEBUG
.
Supply exec_info
to print these.
Sets pipeline_status
and job_status
to failed
on CRITICAL
, supports minor logging levels (41, etc).
To disable the logging on local machines, supply disable_log_formatting=True
when configuring the logger.
Alternatively, without code changes, you can supply the ENV var DISABLE_LOG_FORMATTING"=True
to generally disable it.
Every attribute of the LogRecord will be included in the formatted message unless specified in the excluded_logging_context_keys
list.