Whenever you're ready, walk me through how you'd approach this.
How would you design a distributed key-value store with high availability?
Discuss the consistency model, partitioning, replication, request routing, membership, failure detection, data repair, and operational safeguards. Clarify assumptions before proposing the architecture, and explain how the system behaves during partitions, node failures, replication lag, and recovery.