Debug a cache incident end-to-end

You are paged for elevated latency and a surge in cache misses. Walk through your end-to-end debugging plan: what clarifying questions you ask the on-call/senior partner, which dashboards and metrics you inspect (hit ratio, eviction rate, QPS, p99 latency, connection pools, CPU/memory, network), and how you form and test hypotheses (TTL misconfiguration, capacity pressure, hot keys, thundering herd, hash-ring skew, replication lag, network issues). Propose immediate mitigations and a longer-term fix, then outline post-incident actions (blameless RCA, alerts, runbooks).