Client -> Reverse Proxy -> Upstream Server

It’s quite rare these days for a client (eg web browser, video game, phone app… anything that needs an internet connection, really) to directly connect to the server that is actually serving the request.

It often goes through a reverse proxy which can direct the request to the “best” server (reverse proxies can have multiple upstream servers, either as a cluster that can share the load all dealing with the same service, as a bunch of different services, or both).

This has a lot of benefits for service admins (at the cost of mild complexity), and has near-zero cost for clients.

in a website flow, that could be the server with the site on it or another proxy server.

you would be downsteam from a proxy server and the website its self would be upstream of the proxy server