Fix unscoping of imposed capacity limit by find_each method on Mirror scheduler

What does this MR do?

According to the docs in find_each the method unscopes any LIMIT and/or any ORDER imposed on the list. So what we were doing was finding all the mirrors ready to sync and queuing them all at once. This explains the totally crazy numbers we were seeing.

Does this MR meet the acceptance criteria?

What are the relevant issue numbers?

Merge request reports