Cache Node Failure

One node in the distributed cache cluster has failed. Its keys are redistributing — some concentrate onto the surviving nodes, others fall through as misses that hit the primary database directly. The database must stay within its normal read-latency range: sustained unprotected read pressure against it is treated as a critical incident, not just a cache-layer inconvenience. Restore the cache layer's effective capacity and protect the database from the fallback-read spike before it compounds.

Difficulty

Intermediate

Expected Time

~10 min

Category

Caching