We don't currently clear entries from "status-node-" keyspace when a node is removed from the cluster. This can cause confusion when performing aggregation over those keys. For instance, when working with crdb_internal.kv_store_status and aggregating over metrics->>'ranges.underreplicated'.

I don't know the full history or intention of these keys and the metrics they contain, but we should likely either:

1. clear the metrics portion of the keys associated with decommissioned nodes
2. delete the keys associated with decommissioned nodes entirely

cc. @tbg since you might be one of the few people who still knows how this subsystem works and what its intentions are.