Dumb question, if someone starts off by grouping their data only by Customer ID and then later needs to shard. Couldn’t all of the sharding problems go away if they simply created a new Customer_Tenant table to map customer_id into tenant_group_id
If you have that tenant discriminator on all tables then it's easy to route it to the appropriate physical and logical shard right away vs. having to do some DB join first in your request to figure out where it goes.
Is there a specific term for the trade off of denormalized vs more easily query/shard/etc?
No, it's pretty much just called that. A good DBA will be able to strike the right balance between normalization and performance using their intuition and experience, which is one of the reasons they're often paid very well despite working in relatively ancient ecosystems.
I wonder if this could or should be a built-in feature in databases. It is "meta-data" meaning data about data, who owns it.