In-memory speed for caching, sessions, and queues
Redis is an in-memory data store used for caching, session storage, rate limiting, and lightweight message queuing trading persistence guarantees for extremely fast read and write performance, since data lives in memory rather than on disk.
A lot of application performance work comes down to not hitting the primary database for data that doesn't need a fresh read every time. Redis absorbs that load cached responses, rate-limit counters, session lookups keeping the primary database free for the queries that actually need it.
We use Redis for caching, rate limiting, and session management across the services we build a standard part of the infrastructure layer supporting authenticated, rate-limited APIs described in our security practices.