When I see the insights page for a manage DB instance, I get CPU, memory, load average, and replication status. But no disk IO metrics.

If a disk behind a database instance is getting hammered, I should know about it, as a customer.