looking again, the reporting under 'API Requests' downtime seems accurate but none of the earlier outages report downtime under 'Git Operations', which were also broken.

didn't notice there was a selector for which type of downtime on that page that didn't default to all.