Sentry uses logging's levelname instead of levelno when mapping to levels.

How do you use Sentry?

Self-hosted/on-premise

Version

1.5.12

Steps to Reproduce

Where I work, it has been a convention to do the following in order to get python logging to modify its output to be a little cleaner:

logging.addLevelName(logging.DEBUG, "debug: ")
logging.addLevelName(logging.INFO, "")
logging.addLevelName(logging.WARNING, "warning: ")
logging.addLevelName(logging.ERROR, "error: ")
logging.addLevelName(logging.CRITICAL, "critical: ")
logging.addLevelName(logging.FATAL, "fatal: ")`

This along with:

 _FMT = "%(asctime)s.%(msecs)03d " + SCRIPT_NAME + ": %(levelname)s%(message)s"
LOGGING_FORMATTER = logging.Formatter(_FMT, _DATEFMT)
handler.setFormatter(LOGGING_FORMATTER)

This causes a sentry error that the level name of warning: doesn’t exist. Probably because sentry sdk is mapping to sentry levels using levelname of the log message and not levelno.

Expected Result

Sentry should map from the logging level correctly.

Actual Result

I get an invalid level name of ‘warning:’ in the sentry record in the GUI.

I’ll include my work around below in further comment.

[antonpirker]

This issue is fixed in https://github.com/getsentry/sentry-python/pull/1449 and will be released soon.

Thanks again @rrauenza for your contribution! (See the PR on how to get you some stickers as a thank you!)

[antonpirker]

Hey @rrauenza, I have reviewed your PR: https://github.com/getsentry/sentry-python/pull/1449