Between 1:15 AM and 6:15 AM UTC our API experienced a resource overload incident affecting a small subset of resources that require access to our caching backend. A fix has been deployed and we are actively monitoring the situation. We are working on restoring metadata.