Reverse Proxy — Part 2: Architecture, Production Patterns & Advanced Concepts

maang.io System Design Series

What is this?

In Part 1 the reverse proxy was a concierge — one friendly face at the front desk. Now we go behind the desk and watch what the concierge actually does with each request: how it's read, routed, secured, cached, and handed to the right backend. We'll also finally settle the question everyone gets tangled in — how a reverse proxy differs from a load balancer and an API gateway.

The one-line idea: a production reverse proxy is a small assembly line — each request flows through a pipeline of stages (terminate TLS → parse → route → secure → cache → balance → forward), and most of the engineering is deciding what each stage should do.

Keep the concierge in mind throughout: every stage below is just one more thing the concierge handles so your backends don't have to.