Gunicorn ignores signals

Question

When running gunicorn with --preload flag, the running master worker ignores signals (TTIN, TTOUT, HUP, etc).

gunicorn -b 0.0.0.0:5000 --worker-class eventlet -w 3 --reload server.wsgi:app --graceful-timeout 10 --preload

For example sending TTIN signal kill -TTIN $master_pid does nothing as well as CTRL+C does not interrupt the server, only kills the workers. Is there a reason these commands do not work with --preload flag? Without the flag gunicorn functionality works (signals, ctrl+c) without any problem. Maybe there is a problem with Flask application and not in Gunicorn?

The whole idea behind usage of --preload flag in my case is the scheduler that I use in Flask application. I need to run some tasks in background but if I spawn more than one worker, the code gets forked and executed multiple times across each worker. In my mind when using --preload flag, the scheduler gets initialized and AFTER that workers get forked. With --preload flag scheduler works as expected with multiple workers but I am stuck with original problem (ignoring signals, etc.)

EDIT: the problem seems to be with using monkey_patch of eventlet with --preload (github) I am using in my application. Steps to reproduce the issue:

flask_app.py

  1 import eventlet
  2 eventlet.monkey_patch()
  3 from flask import Flask
  4
  5 app = Flask(__name__)
  6
  7
  8 @app.route("/")
  9 def hello():
 10     return "<h1 style='color:blue'>Hello There!</h1>"
 11
 12 if __name__ == "__main__":
 13     pass

wsgi.py

  1 from flask_app import app
  2
  3 if __name__ == "__main__":
  4     app.run()

command to start gunicorn server:

gunicorn -b 0.0.0.0:5000 --worker-class eventlet -w 3 wsgi:app --preload