Kanban board - IssueIdCallback loads documents to get only id

Problem

Loading of Kanban is slow in case of many Swimlanes and/or CardColors.
See kanban-board loading is slow for context.
This computation is done as part of /rest/greenhopper/1.0/xboard/work/allData.json API request.
One of the reasons: IssueIdCallback is loading the whole Lucene document to extract only issue_id.
Context: This callback is used for swimlanes, card colours and getting ids of visible issues.

Suggestion

It could be rewritten to operate with SortedDocValues, which should be faster since it doesn't require to load the whole document from the index

Results

Ridiculous dataset:

• Backlog with 10k issues
• 11 JQL-based colours
  • Each color query returns almost all 10k issues

Time to calculate colours:
1,4sec -> 80msec