It's getting increasingly difficult to manage datastores without knowing what cluster they are associated with. Major pain points:

• when connecting to the datastore, we need to look for the secret in the correct cluster that it was deployed in when the datastore was created (and the one it was auto-vpc-peered to)
• when allowing users to connect apps to the datastore (i.e., attaching the datastore env group to them), the apps need to be deployed in the cluster that the env group was created in

I see this field as a necessary evil which will be deprecated once we decouple datastores from clusters