A GitHub outage caused some of our API endpoints started to fail. To reduce the impact we briefly turned off some communication with GitHub which made things stabilize on our side. At this time, no Happo statuses could be posted to GitHub PRs and commits. At 18.11 UTC we saw things starting to stabilize and briefly after we were back in a good state again.

We’re seeing errors coming from our own API calls to GitHub causing most jobs to fail on our end as well. We’re investigating this right now.