notch 2025.1

The Python logging micropackage

notch

🪓 The Python logging micropackage

Purpose

You can use notch to quickly configure logging for your Python application. I designed notch to encapsulate and standardize logging configuration for Python applications that are deployed as container workloads. The default settings work in many simple cases, and you can further customize logging behavior with environment variables.

Usage

Import notch and call configure() to configure logging in your application entrypoint.

import notch

notch.configure()

Call configure() as early as possible in your application startup, because notch in turn calls logging.basicConfig() to configure the logging system.

Using the default settings, notch will configure the root logger to send logs to sys.stdout, set the log level to INFO, and set the log format to '%(levelname)s [%(name)s] %(message)s'.

Configuration

You can change the behavior of notch by setting environment variables.

• LOG_LEVEL: change the log level for the root logger. This can be any level the Python logging module recognizes: DEBUG, INFO (the default), WARNING, ERROR, or CRITICAL.
• LOG_FORMAT: change the format for log output. Use any placeholders defined in the Python logging documentation.
• OTHER_LOG_LEVELS: change the log level for loggers other than the root logger. See the explanation below.

Setting log levels for other modules

If you use third-party packages, you may want to set log levels for these packages independently. For example, you may want to see DEBUG logs for your own code, but only WARNING messages for other modules. You can do this by setting the OTHER_LOG_LEVELS environment variable.

The value of this variable should be a module name and a log level separated by a colon. You can set log levels for multiple modules by separating entries with a space.

OTHER_LOG_LEVELS=first_module:WARNING second_module:ERROR

For a concrete example, if your application depends on boto3 but you only want to see errors from boto3-related modules in your logs, configure like this:

OTHER_LOG_LEVELS=boto3:ERROR botocore:ERROR

You can specify submodules as well if you need to be more specific:

OTHER_LOG_LEVELS=apscheduler.scheduler:WARNING