Certain pages (e.g. the group wide issues page) filter miletones by their title. Without an index this will result in a sequence scan on a large dataset increasing the total loading time of a page.
