We’ve shared a technical write-up about the causes of this incident and steps we’ve taken to improve reliability here: https://incident.io/blog/intermittent-downtime