Fixing NLS_SORT for Jira Agile queries

JAG needs a better way of fixing the NSL_SORT in Oracle DB.
Right now we are doing a forced NSL_SORT over the RANK column dropping the performance.

Possibilities:

1. Document this
2. Do a migration script that force the database to NLS_SORT=Binary
3. Try to fix the index or the column to use the proper NLS_SORT.

The difference between this ticket and GHS-11162 is that we should take off the NLS_SORT while the other is about removing the function from the code.