Caching strategies for high-traffic APIs

hey everyone, i'm working on a project with an api that gets hit with tons of read traffic. we've put in some basic caching, but i'm curious about better ways to do it. has anyone had luck with stuff like cache invalidation patterns, you know, write-through or write-behind, or maybe distributed caching solutions that are more than just in-memory? what do you guys usually do to keep api responses up-to-date without killing the backend