baekenough/redis-best-practices

Redis Best Practices

Caching Patterns

Cache-Aside (CRITICAL)

• Read: check cache → miss → read DB → set cache
• Write: update DB → invalidate cache
• Best for read-heavy workloads

Write-Through

• Write: update cache AND DB simultaneously
• Ensures consistency
• Higher write latency

Write-Behind

• Write: update cache → async DB update
• Lowest latency
• Risk of data loss

Data Structures

String

• Simple key-value: SET key value
• Counters: INCR, DECR, INCRBY
• Bit operations: SETBIT, BITCOUNT

Hash

• Object storage: HSET user:1 name "Alice"
• Partial updates: HINCRBY user:1 visits 1
• Memory efficient for small hashes

List

• Queues: LPUSH, RPOP (FIFO)
• Stacks: LPUSH, LPOP (LIFO)
• Capped collections: LTRIM

Set

• Unique collections: SADD, SMEMBERS
• Intersections: SINTER
• Random sampling: SRANDMEMBER

Sorted Set

• Leaderboards: ZADD, ZRANGEBYSCORE
• Rate limiting: ZADD timestamp
• Priority queues

Stream

• Event log: XADD, XREAD
• Consumer groups: XGROUP, XACK
• Pub/Sub with persistence

Performance

Memory Optimization

• Set maxmemory policy: allkeys-lru, volatile-lru
• Monitor memory: INFO memory
• Use appropriate data structures (Hash vs String)

Pipelining

• Batch multiple commands
• Reduces round-trip latency
• Use for bulk operations

High Availability

Redis Cluster

• Horizontal scaling
• Automatic sharding (16384 slots)
• Multi-master replication

Redis Sentinel

• Automatic failover
• Monitoring and notifications
• Configuration provider

References

• Redis Best Practices