Currently, we do 2 separate queries to compute excess storage for
projects with project-level limits and those that inherit the global
limit.

This improves the calculation so that it does both in one query.