The stack trace is missing when an exception is thrown.

Hi,
I'm using datadog and collect span with some intrumentation (httpx, wsgi), but when an exception is raised, the stacktrace is missing in datadog.
I looked at the instrumentation code and saw that the command span.record_exception(exception) was missing here:

https://github.com/open-telemetry/opentelemetry-python-contrib/blob/5a15f57588fa94e3f761c1bd9400ab5a6a66a9c8/instrumentation/opentelemetry-instrumentation-httpx/src/opentelemetry/instrumentation/httpx/__init__.py#L875

            if exception:
                if span.is_recording() and _report_new(
                    self._sem_conv_opt_in_mode
                ):
                    span.set_attribute(
                        ERROR_TYPE, type(exception).__qualname__
                    )
                    metric_attributes[ERROR_TYPE] = type(
                        exception
                    ).__qualname__

                raise exception.with_traceback(exception.__traceback__)

What do you think? I'm not a coding expert, but if it seems relevant to you, I'd be happy to submit a pull request.

Thanks a lot,

Axel

Start in instrumentation/opentelemetry-instrumentation-httpx/src/opentelemetry/instrumentation/httpx/init.py around the exception-handling code at the linked line. Review how the command span handles the exception and verify the related instrumentation tests, if present. Done means exceptions raised there produce a stack trace in the collected Datadog span.