The single-tenant cache
Why we moved our caching layer from multi-tenant to single-tenant, what that mental model unlocks for latency and operations, and how we're shaping the cache as a leaderless network of nodes that lives close to wherever you happen to be building.