From 4:07 to 5:18 UTC we experienced elevated error rates for webhook, UI and API requests for customers on shared stacks. The issue is now resolved. Non-webhook action runs were unaffected, and customers on dedicated stacks were unaffected.