What is the rationale behind this?
By putting a proxy in front of the port, you are not increasing the protection.
You are doubling the vulnerability, because now there are 2 software components whose (theoretically possible) vulnerabilities may be exploited by an attacker.
So far there have been no known exploits of the exposed port of a Plex Media Server.
I have several reasons, but I am far from being an expert, so you might debunk my ideas.
My two main reasons are:
masking/protecting the backend
easy and automated issuing of SSL certificates
The reverse proxy service exists anyway, so it would be unnecessary not to use it and instead create an isolated solution for plex where I have to e.g. manually renew the SSL certificate etc.
But after some reading I think I understood anyway that the message is not correct and that connecting, for example, a TV to the domain that redirects to Plex would be a direct connection?
Services like Shodan will be able to identify Plex anyway. No matter if it’s behind a proxy or not.
A client-server connection is considered “local” if both are in the same local network segment. Which has nothing to do with domain names. Only with IP adresses and net masks.