persistent sessions

Persistent sessions maintain user state across server restarts and browser closures by offloading session data from volatile RAM to durable storage like Redis or PostgreSQL.

Stateless protocols like HTTP require a reliable mechanism to track users. Persistent sessions solve this by mapping a unique browser cookie to a backend database entry (often a key-value store like Redis). This architecture prevents 'session drops' during high-traffic deployments or server failures. By setting an explicit Time To Live (TTL) of 30 days or more, platforms like GitHub or Amazon ensure users remain authenticated even after a full system reboot. It shifts the burden of state from the local application memory to a distributed layer, allowing for seamless horizontal scaling across multiple availability zones.

https://redis.io/glossary/session-persistence/