gunicorn keeps booting workers when exceptions raised at startup events

Hi, I have the same issues when deploy FastAPI using gunicorn + uvicorn，it keeps booting workers when exceptions raised at startup events.

main.py

from fastapi import FastAPI

app = FastAPI()


@app.on_event("startup")
async def startup():
    raise Exception("error")


@app.get("/")
async def hello():
    return {"msg": "hello world"}

gunicorn.conf.py

import json
import os

worker_class = "uvicorn.workers.UvicornWorker"

host = os.getenv("HOST", "0.0.0.0")
port = os.getenv("PORT", "8000")
bind_env = os.getenv("BIND", None)
use_loglevel = os.getenv("LOG_LEVEL", "info")
if bind_env:
    use_bind = bind_env
else:
    use_bind = f"{host}:{port}"


accesslog_var = os.getenv("ACCESS_LOG", "-")
use_accesslog = accesslog_var or None
errorlog_var = os.getenv("ERROR_LOG", "-")
use_errorlog = errorlog_var or None
graceful_timeout_str = os.getenv("GRACEFUL_TIMEOUT", "120")
timeout_str = os.getenv("TIMEOUT", "120")
keepalive_str = os.getenv("KEEP_ALIVE", "5")

# Gunicorn config variables
loglevel = use_loglevel
bind = use_bind
errorlog = use_errorlog
accesslog = use_accesslog
graceful_timeout = int(graceful_timeout_str)
timeout = int(timeout_str)
keepalive = int(keepalive_str)

# For debugging and testing
log_data = {
    "loglevel": loglevel,
    "bind": bind,
    "graceful_timeout": graceful_timeout,
    "timeout": timeout,
    "keepalive": keepalive,
    "errorlog": errorlog,
    "accesslog": accesslog,
}
print(json.dumps(log_data))

gunicorn log

{"loglevel": "info", "bind": "0.0.0.0:8000", "graceful_timeout": 120, "timeout": 120, "keepalive": 5, "errorlog": "-", "accesslog": "-", "host": "0.0.0.0", "port": "8000"}
[2021-06-01 21:13:05 +0800] [40946] [INFO] Starting gunicorn 20.0.4
[2021-06-01 21:13:05 +0800] [40946] [INFO] Listening at: http://0.0.0.0:8000 (40946)
[2021-06-01 21:13:05 +0800] [40946] [INFO] Using worker: uvicorn.workers.UvicornWorker
[2021-06-01 21:13:05 +0800] [40949] [INFO] Booting worker with pid: 40949
[2021-06-01 21:13:05 +0800] [40949] [INFO] Started server process [40949]
[2021-06-01 21:13:05 +0800] [40949] [INFO] Waiting for application startup.
[2021-06-01 21:13:05 +0800] [40949] [ERROR] Traceback (most recent call last):
  File "./envs/test/lib/python3.7/site-packages/starlette/routing.py", line 526, in lifespan
    async for item in self.lifespan_context(app):
  File "./envs/test/lib/python3.7/site-packages/starlette/routing.py", line 467, in default_lifespan
    await self.startup()
  File "./envs/test/lib/python3.7/site-packages/starlette/routing.py", line 502, in startup
    await handler()
  File "./app/main.py", line 16, in startup
    raise Exception("error")
Exception: error

[2021-06-01 21:13:05 +0800] [40949] [ERROR] Application startup failed. Exiting.
[2021-06-01 21:13:05 +0800] [40949] [INFO] Worker exiting (pid: 40949)
[2021-06-01 21:13:05 +0800] [40950] [INFO] Booting worker with pid: 40950
[2021-06-01 21:13:06 +0800] [40950] [INFO] Started server process [40950]
[2021-06-01 21:13:06 +0800] [40950] [INFO] Waiting for application startup.
[2021-06-01 21:13:06 +0800] [40950] [ERROR] Traceback (most recent call last):
  File "./envs/test/lib/python3.7/site-packages/starlette/routing.py", line 526, in lifespan
    async for item in self.lifespan_context(app):
  File "./envs/test/lib/python3.7/site-packages/starlette/routing.py", line 467, in default_lifespan
    await self.startup()
  File "./envs/test/lib/python3.7/site-packages/starlette/routing.py", line 502, in startup
    await handler()
  File "./app/main.py", line 16, in startup
    raise Exception("error")
Exception: error

[2021-06-01 21:13:06 +0800] [40950] [ERROR] Application startup failed. Exiting.

I hope the gunicorn would exit when exceptions raised, but it keeps booting workers. Can someone helps me, thanks?

_Originally posted by @fanchunke1991 in https://github.com/encode/uvicorn/issues/1031#issuecomment-852121760_

[Kludex]

Oh. I think you’ve confused the current issue. This issue is for gunicorn exit behavior when startup fails. We don’t have a matching behavior for uvicorn yet.

We have this PR from some long time ago: https://github.com/encode/uvicorn/pull/835 Which I don’t think we should merge, as I’d prefer to match the exit code from gunicorn worker.

[vackosar]

@Kludex thanks, yes. Any non zero exit code there would be great (even exit code 3 like you suggest here if I understand well!). Otherwise, how to find out that the app failed? Returning 0 is destroying information for me. Even if it required setting a switch or running through gunicorn, just to have some way to find out that it failed on startup.