1👍
✅
While it is possible to have such a setup (and it was pretty common some years ago under uWSGI) it causes (soon or later) headaches all over the place.
The best (or better: saner) approach is starting a single process for each uWSGI instance and let it scale to more workers when needed:
–processes 8 –cheaper 1 –idle 60
will start each instance with only one worker, and eventually (as needed) will raise up to 8. In addition to this when the instance is idle for more than 60 seconds it will be put on hold (only the master running) until another request is issued.
There are other fine-tuning you can make, but the general rule is that multiple interpreters under django (for the way it works) are not an easy path to follow.
Source:stackexchange.com