autocat3 is using the built-in CherryPy HTTP server which is multi-process but single-threaded. This is resulting in what appear to be python GIL limits in production at higher thread counts limiting overall app performance.
We need to evaluate moving CherryPy to a multi-process wsgi server that won't have these limitations, eg: https://docs.cherrypy.dev/en/latest/deploy.html#wsgi-servers
gunicorn with its gthread worker sounds like the best candidate as it's both multi-process and multi-threaded: https://gunicorn.org/design/#__tabbed_1_2
autocat3 is using the built-in CherryPy HTTP server which is multi-process but single-threaded. This is resulting in what appear to be python GIL limits in production at higher thread counts limiting overall app performance.
We need to evaluate moving CherryPy to a multi-process wsgi server that won't have these limitations, eg: https://docs.cherrypy.dev/en/latest/deploy.html#wsgi-servers
gunicorn with its gthread worker sounds like the best candidate as it's both multi-process and multi-threaded: https://gunicorn.org/design/#__tabbed_1_2