We've concluded our investigation into the incident. It was caused by an unexpected issue with system resource allocation on our API infrastructure, which has now been fully resolved.

API recovered.

API went down.