Numbers of calls in FastAPI initialization confuses me

Question

For example, let's take simple FastAPI app:

import uvicorn
from fastapi import FastAPI

print("calling main!")
app = FastAPI()

@app.get('/test', status_code=200)
async def test():
    return True


if __name__ == "__main__":
    uvicorn.run("main:app", use_colors=True, workers=2)

If we run it, we get following output:

calling main!
INFO:     Uvicorn running on http://0.0.0.0:8020 (Press CTRL+C to quit)
INFO:     Started parent process [143621]
calling main!
calling main!
INFO:     Started server process [143832]
INFO:     Waiting for application startup.
INFO:     Application startup complete.
INFO:     Started server process [143829]
INFO:     Waiting for application startup.
INFO:     Application startup complete.
calling main!
calling main!

So, for app with two workers (parent + 2 server process) we have 5 (!) execution of main. Why? Maybe I'm missing something simple, but it confuses me. Thanks in advance for answers!

Answer 1

Based on this answer about miltiprocessing.

Uvicorn start new processes to create workers. It will import parent, that gives you 1 "calling main!" for each worker. Than the worker will start new server, which gives you 1 more "calling main!" for each. worker.

You can try print('worker pid: {}, module name: {}'.format(os.getpid(), __name__)) instead of your print to check processes.