if one of the databases’ usage is resources/io heavy we can move it to its own dedicated hardware without any impact on the clients.
If hardware isolation is your concern, you can get this in CockroachDB using zone configurations. You can assign attributes to your nodes and then add a constraint to a database or table that causes it to be replicated only to nodes with matching attributes.
Are there plans for a 3 tier namespace similar to Postgresql? I did not see mention of such plans in the RFC
An earlier draft of the RFC proposed reworking things to use a three-level namespace, but it was too late to make breaking changes as we were stabilizing for 1.0. Now, we’re still vaguely interested in moving towards a three-level namespace, but we’ll have to figure out how to do it without breaking existing applications, and we don’t have any plans to work on this in the near future.