Sort burnup data chronologically after merging synthetic and "real" data

Summary:
Ref T13020. See PHI273. See D18853. On secure, the chart looks less promising than it did locally, and is full of discontinuities:

I think this is a sorting issue. But if I can't fake my way through this soon I'll maybe get the Fact engine running and use it to provide the data here, as a sort of half-step toward T1562?

Test Plan: Chart looks the same locally, will push and see if secure improves?

Reviewers: amckinley

Reviewed By: amckinley

Maniphest Tasks: T13020

Differential Revision: https://secure.phabricator.com/D18854