3 · Caching · Eviction & failure modes

Eviction & failure modes

LRU eviction
Drop least-recently-used keys when full. The usual default.
TTL
Expire keys after N seconds — a safety net that bounds staleness.
Thundering herd / stampede
A hot key expires; thousands of requests miss at once and hammer the DB. Fix: request coalescing/lock, jittered TTLs, serve stale while refreshing.
Hot key
One key gets huge traffic (a celebrity). Fix: local in-process cache, replicate the key.
Cache penetration
Requests for keys that don't exist always miss. Fix: cache negative results, bloom filter.

1. Why delete the cache key on update instead of writing the new value?

2. A homepage key expires and DB CPU spikes to 100%. Name for this?

Answered 0 / 2

Explain When should you NOT add a cache?