FastAPI Python Web Service Deploying For Hours

[auhide]

Hello,

I’ve started the deployment of a FastAPI Python app, but it doesn’t seem to finish.
The logs say that the server is up and running:

I can also make calls to the API but a response never seems to come, the request is processed infinitely.

What could the problem be? If you need any more details, I’ll be glad to provide!

Best Regards,
Adam

[splatcollision]

I ran into this a few months ago and solved it like this:

I use a shell script that runs uvicorn for my FastAPI app on render. You have to use port 10000 AND 0.0.0.0 for host, which isn’t documented anywhere but some examples.

Here’s a shell script that can be used to start your server - invoke this as your ‘start command’ in render config.

#!/bin/bash
# my app is in main.py and the FastAPI class instance is named 'app'
uvicorn main:app --host 0.0.0.0 --port 10000

[auhide]

Thank you, @splatcollision!! Defining the certain host and port you’ve mentioned, worked!

It should really be mentioned in the documentation. I thought that the default port for gunicorn (8000) would be the one that’s expected from Render, since in the Flask example the deployment command was gunicorn without any arguments.