CostCategory model with a name, description, source_type, system_default, namespace, and label fields

• Unique constraint on name 
• The namespace column should be an array field of namespace names
• Optional: The label column should be an array field of label keys – maybe values??? - fortunately we don’t need to quite decide the label handling for this epic, would be design work for future customer defined categories
• We also need to account for a wildcard for LIKE matching, at least for this platform cost bucket because some clusters will have different sets of kube- and openshift- projects
  • Example for platform cost: namespace column: [“openshift-%”, “kube-%”, “Platform Unallocated Capacity”]}

Are there additional constraints? E.g. name must be unique.  What indices should we provide and what is the default order_by for this model.