A hot key concentrates traffic on one shard.
Local in-process caching (LRU cache in the app layer) with short TTL
Key hashing with suffix (product:1:{shard_0}, product:1:{shard_1}) to spread across slots — but loses atomic ops
Read replicas with READONLY mode in Cluster for read-heavy hot keys
Proxy-level fan-out (Twemproxy, Envoy)
0-2 years experience
2-5 years experience
5-8 years experience
8+ years experience
