How to tune parameters max-requests and max-requests-jitter?

[slava-pasedko]

I have a production Django environment which is being served by gunicorn.

I start an application using this command:
gunicorn config.wsgi:application -w 80 -b 0.0.0.0:8000 --keep-alive 30 -t 120

What characteristics of my production environment I should know to calculate the right amount of max-requests and max-requests jitter? RPS? CPU cores? etc.

I've seen articles that sometimes after people add this options with small values they have 500 on their production environment.

The question is how to calculate the right amount of requests and how to test it?
What tools can I use to simulate the production environment load in order to understand that those options help me to reduce the amount of RAM used by the cluster?