We had a short incident just now that affected API availability related to a failed deployment. The change was immediately rolled back, and it is unlikely that users noticed the downtime.