Resolved
Predictions queues are back into normal range and the affected systems remain stable. Thank you for your patience!
Monitoring
Autoscaling capabilities were nonfunctional, which resulted in cold deployments not scaling out and warm deployments not scaling appropriately based on demand.
