Document multi-worker, pre-fork server metrics limitation in Troubleshooting for Zero-Code Python

[tammy-baylis-swi]

What needs to be changed?

A commonly-encountered limitation of Zero-Code Python instrumentation is that pre-fork servers set up with multiple workers often break metrics generation. It would be helpful if this was documented in the new Troubleshooting page introduced with #6541.

What is the name + path of the page that needs changed?

Troubleshooting for Zero-Code Python: content/en/docs/zero-code/python/troubleshooting.md

[svrnm]

Thanks for raising this @tammy-baylis-swi!

@open-telemetry/python-approvers PTAL

[emdneto]

I think it would be nice to have a source of truth for this. We receive questions about it frequently. I was thinking on a compliance matrix showing what Works and what doesn't works

Something like:

Stack     |  Traces | Metrics | Logs |
Uvicorn |.    V.     |.    X.       |.    V  |
Gunicorn |.    V.     |.    X.       |.    V  |
Gunicorn + Uvicorn workers | ...