Caching Strategies with Redis

Cache-Aside (Lazy Loading)

1. Ask Redis for the key.
2. On miss, load from the database.
3. Write the value back with a TTL.
4. Serve subsequent reads from Redis.

TTL expires stale entries; the app is the source of truth, so the DB stays consistent.

TTL Policies

CONFIG GET maxmemory-policy
# noeviction | allkeys-lru | volatile-lru | allkeys-lfu

With a maxmemory cap, eviction preserves the hottest keys (LRU/LFU) instead of failing writes.

Cache Consistency

Invalidate on writes: after updating the DB, delete or update the Redis key so readers never serve stale data.

Use It for

• Session storage with TTL.
• Hot query results and API responses.
• Rate limit counters (INCR + EXPIRE).
• Leaderboards and ranking caches.

Key Points

• Cache misses flow to the DB and fill Redis.
• TTL and eviction bound memory and staleness.
• Write-through invalidation prevents staleness.
• Sessions, counters and responses all fit the model.