I was looking at the setup of reverse proxies, like nginx in front of apache, to reduce load on webservers for serving static content.
For any given URL, our web pages will have different parts static and dynamic. How do reverse proxies know when to serve resources themselves, or pull it from the source web server behind them? Obviously things like images and javascript files are static, but even some of those, like captchas, are dynamically generated. Our drupal sites actually look at the path in the url and serve different content accordingly, taking path elements as arguments.
Does it simply require a lot of configuration on the reverse-proxy side, with the configurator having deep knowledge about which parts of the site are dynamic versus static?