Seeking clarification for endpoint configuration behind an nginx load balancer

I’m trying to wrap my head around the endpoint configuration, especially in prod settings.

Generally I have something like this:

config :btest, BtestWeb.Endpoint,
    url: [host: "localhost", port: 4480, scheme: "http"],
    http: [
      # Enable IPv6 and bind on all interfaces.
      # Set it to  {0, 0, 0, 0, 0, 0, 0, 1} for local network only access.
      # See the documentation on https://hexdocs.pm/bandit/Bandit.html#t:options/0
      # for details about using IPv6 vs IPv4 and loopback vs public addresses.
      ip: {127, 0, 0, 1},
      port: port
    ],

The idea is to have the Phoenix app behind an nginx load balancer, offloading the SSL handshake to it, I’d say not a very unusual set up if you want to run more than one app on the same server. I do care only about binding to localhost on a certain port and that’s that.

What I can’t understand: I can set the port variable to a different value than the one hardcoded in the url , let’s say 4481. The server goes up, but the messages are a little bit confusing:

[info] Running BtestWeb.Endpoint with Bandit 1.6.10 at 127.0.0.1:4481 (http)
[info] Access BtestWeb.Endpoint at http://localhost:4480

Actually, it listens on 4481 only.
So, my questions are:

• why do I need the port two times, and what parts of the Endpoint are configured via these two options? (I tried to look in the code with no success)
• In which specific situation I would need to provide a different value for the port in http and the one configured via the env variable?