Using additional command line arguments with gunicorn
You can't pass command line arguments directly but you can choose application configurations easily enough.
$ gunicorn 'mypackage:build_app(foo="bar")'
Will call the function "build_app" passing the foo="bar" kwarg as expected. This function should then return the WSGI callable that'll be used.
I usually put it in __init.py__
after main()
and then I can run with or without gunicorn (assuming your main()
supports other functions too).
# __init__.py# Normal entry pointdef main(): ...# Gunicorn entry point generatordef app(*args, **kwargs): # Gunicorn CLI args are useless. # https://stackoverflow.com/questions/8495367/ # # Start the application in modified environment. # https://stackoverflow.com/questions/18668947/ # import sys sys.argv = ['--gunicorn'] for k in kwargs: sys.argv.append("--" + k) sys.argv.append(kwargs[k]) return main()
That way you can run simply with e.g.
gunicorn 'app(foo=bar)' ...
and your main()
can use standard code that expects the arguments in sys.argv
.