Top 35 Redis Interview Questions & Answers 2025

Strings: simple key-value, counters, caching. Lists: queues, recent items. Sets: unique items, tags, relationships. Sorted Sets: leaderboards, rate limiting, priority queues. Hashes: object storage, user sessions. Streams: event sourcing, message queues. HyperLogLog: cardinality estimation. Bitmaps: flags, real-time analytics.

RDB (snapshots): point-in-time snapshots, compact files, faster restarts, potential data loss between snapshots. AOF (append-only file): logs every write, configurable fsync (always, every second, never), larger files, slower restart. Best practice: use both - AOF for durability, RDB for backups/faster recovery. Configure based on data loss tolerance.

Strategies: TTL expiration (set expire time), active invalidation (delete on update), write-through (update cache with DB), cache-aside (application manages), pub/sub for distributed invalidation. Patterns: cache stampede prevention (locking, probabilistic early expiration), lazy loading, refresh-ahead. Key design: include version in key for easy invalidation.

Redis Cluster: automatic sharding across nodes using hash slots (16384 slots). Each master handles subset of slots with replica(s). Failover: replicas monitor master via heartbeat, elect new master using Raft-like consensus when master fails. Client redirects: MOVED (permanent), ASK (temporary during resharding). Minimum 3 masters recommended for quorum.

MULTI/EXEC: batch commands atomically. WATCH: optimistic locking - abort if watched keys change. Limitations: no rollback on error (commands after error still execute), all-or-nothing queuing only. For complex transactions: Lua scripts (atomic execution, reduce round trips). Redis is single-threaded, so individual commands are atomic.