Something seriously wrong when running reflex app in docker container

[NanoscaleSimulations]

Hello there, 😊

There is something wrong when running a reflex app in a docker container. The front-end works fine, but the backend tends to work at its own free will.
I am using the exact dockerfile from the example and the same rxconfigs. Sometimes the backend will work, but for the most time, it will not.
I have even tried to play around with the api_url parameter in rxconfig and I even played with changing the ports.

Can anyone explain to me a fix for this?

Kind regards
Zan

Ps. I am trying to run it on a Oracle Cloud Instance (Linux/arm64)

[NanoscaleSimulations]

So i figured out myself that i need to set my api_url to not only the backend port (in my case 8000), but also to the ip of my cloud instance 😂🔫. This has been a very confusing struggle for me. I had a hunch that something was wrong with my config, but due to lack of descriptive documentation about what api_url actual does, it took long to figure out.
However, for anyone with the same problem, the solution is:

In rxconfig.py:
api_url = "http://serveripaddress:port"

Even when using docker 😅