[WSGI] waitress wg. gunicorn

Question about the WSGI story of Zope 4/Plone 5.2: the current default setup for running Plone under WSGI uses Waitress. Would it be possible to use a different WSGI middleware like Gunicorn? Would be gain something from using Guicorn like being able to start N instances of the service, restarting an instance after N requests etc.? In general N instances running under control of one Gunicorn server would share the same instance home...unlikely to work!? Thoughts?

I don't know any about gunicorn or waitress.
For my Django work I'm using uWSGI behind nginx with uwsgi_pass. I've configured https://uwsgi-docs.readthedocs.io/en/latest/Cheaper.html to start 2 workers initially and spawn more as needed.
One could use https://uwsgi-docs.readthedocs.io/en/latest/Options.html#max-requests to auto-restart workers.

Experience from others is more than welcome :smiley:

The question was not about a particular WSGI framework about the chance to run multiple copies of the same Plone instance installation under control of e.g. Gunicorn...clearly this will not work with a standalone ZODB setup because of the lockfile but this could work with ZEO setup.

Plone Foundation Code of Conduct