Enhancing Error Troubleshooting by Adding Traceback to Exception Logs

[pythrick]

Hello,
I've been working with the code and noticed an opportunity to improve our error logging and debugging experience.

Current Code Handling:

Below is how our current code handles and logs exceptions:

Propan/propan/brokers/_model/broker_usecase.py

Lines 614 to 642 in fd243da

	def _log_execution( # type: ignore[override]
	self,
	func: Callable[
	[PropanMessage[MsgType]],
	Awaitable[T_HandlerReturn],
	],
	**broker_args: AnyDict,
	) -> Callable[[PropanMessage[MsgType]], Awaitable[T_HandlerReturn]]:
	@wraps(func)
	async def log_wrapper(message: PropanMessage[MsgType]) -> T_HandlerReturn:
	log_context = self._get_log_context(message=message, **broker_args)
	with context.scope("log_context", log_context):
	self._log("Received", extra=log_context)
	try:
	r = await func(message)
	except SkipMessage as e:
	self._log("Skipped", extra=log_context)
	raise e
	except Exception as e:
	self._log(repr(e), logging.ERROR)
	raise e
	else:
	self._log("Processed", extra=log_context)
	return r
	return log_wrapper

In its current state, the code logs exceptions but doesn't include a traceback or the log_context that could potentially be very useful for debugging.

My Proposal:

To enhance our error logging, I suggest adding the traceback to exception logs, and to provide extra context, include log_context. Here's a potential way of doing it:

# ...
                try:
                    r = await func(message)
                except SkipMessage as e:
                    self._log("Skipped", extra=log_context)
                    raise e
                except Exception as e:
                    self._log(repr(e), logging.ERROR, extra=log_context, exc_info=e)
                    raise e
# ...
    def _log(
        self,
        message: str,
        log_level: Optional[int] = None,
        extra: Optional[AnyDict] = None,
        exc_info: BaseException | bool | None = None,
    ) -> None:
        if self.logger is not None:
            self.logger.log(
                level=(log_level or self.log_level), msg=message, extra=extra, exc_info=exc_info
            )

By incorporating these, we could have better insights into errors which could streamline our debugging process.

Looking for Suggestions:

The part I am pondering over is how to allow users to opt in or out of this additional logging behavior, maybe using an environment variable.

I would love to hear your thoughts and any potential solutions for this part. Thank you in advance for your input.

[Lancetnik]

It is just a decision I use in a new major draft😅
Can u make a PR to implement it here too?

[Lancetnik]

So, you can find it in 0.1.5.14

[pythrick]

Thanks!