Handling Hot Keys
What happens when Cristiano Ronaldo posts to his 650+ million Instagram followers?
Within seconds, millions of users try to view the same post. The cache node holding that post gets hammered with requests while other nodes sit idle. The single Redis server responsible for that key becomes the bottleneck, and response times spike across the entire platform.
This is the hot key problem, and it's one of the most common ways distributed systems fail in production. You can have a perfectly designed, horizontally scaled system with 100 cache nodes, and a single hot key can reduce your effective capacity to what one node can handle.
The problem is deceptively simple to understand but surprisingly difficult to solve. Hot keys appear everywhere: a viral tweet, a flash sale product, a breaking news article, a popular live stream, a celebrity's profile page. Any time traffic concentrates on a single piece of data, you have a hot key waiting to take down your system.
What makes hot keys particularly dangerous is that they often strike without warning. A tweet that was ordinary five minutes ago can become the center of a global conversation. A product that saw normal traffic all year can suddenly receive a million concurrent requests during a flash sale. Your system needs to handle these situations gracefully, not just survive them.
Handling hot keys is a critical skill in system design interviews. Whether you're designing Twitter, an e-commerce platform, a ticketing system, or a live streaming service, the interviewer will expect you to identify and address this challenge.