https://opentelemetry-python.readthedocs.io/en/latest/examples/fork-process-model/README.html indicates that BatchSpanProcessor does not support forking process models, including those used by Gunicorn and uWSGI. However those docs (#1609) pre-date other changes which appear to have added support for forking process models, including:

• Make batch processor fork aware and reinit when needed #2242, and
• bugfix: batch processor doesn't work with uwsgi #2277

which seem specifically targetted at resolving the described issues.

What is the current expected state of support for the various forking models?
Either way, updating the docs to clarify the actual situation would be appreciated.

Possibly related issues:

• Fork process models for metric batch reader do not work #2767
• (exporter-jaeger-thrift): Exception in forked environments #2837
• Out of the box support for uWSGI opentelemetry-python-contrib#675
• Out of the box support for Gunicorn opentelemetry-python-contrib#676