[Python] How to use workers with hypercorn serve?

I want to start my FastAPI app using hypercorn from main.py.

I used this command to start the server manually:

hypercorn --access-logformat "%(R)s %(s)s %(st)s %(D)s %({Header}o)s" --access-logfile - -b 0.0.0.0:8000 -w 9 app:app

Now I need to do the same thing from my code. First attempt was:

from hypercorn.config import Config
from hypercorn.asyncio import serve

from api.app import app # app = FastAPI()

config = Config()
config.bind = ['0.0.0.0:8000']
config.access_log_format = '%(R)s %(s)s %(st)s %(D)s %({Header}o)s'
config.accesslog = mainLog # mainLog = logging.getLogger('main')
config.loglevel = 'INFO'

asyncio.run(serve(app, config))


I've tried to set config.workers = 9 , but in response I've got:

Warning: The config `workers` has no affect when using serve
warnings.warn("The config `workers` has no affect when using serve", Warning)


Then I rewrite my code to this:

import logging
import asyncio

from hypercorn.config import Config
from hypercorn.asyncio import serve

from api.app import app


mainLog = logging.getLogger('main')
...

async def main(app, config):
await serve(app, config)

if __name__ == "__main__":
config = Config()
config.bind = ['0.0.0.0:8000']
config.access_log_format = "%(R)s %(s)s %(st)s %(D)s %({Header}o)s"
config.accesslog = mainLog
config.loglevel = 'INFO'
config.worker_class = 'asyncio'
config.workers = 9
try:
loop = asyncio.get_event_loop()
loop.run_until_complete(main(app, config))

except:
mainLog.exception('Fail to start server')


But I still got the same warning. Documentation says:

... it is possible to use Hypercorn programmatically via the serve function available for either the asyncio or trio Workers (note the asyncio serve can be used with uvloop).
​

Am I misunderstand the docs? I have lack of knowledge in this area, so please help me.

Continue reading...