We are currently investigating this issue.

EDIT: We're seeing service stability return. We're continuing to work towards full resolution.