Customers can query workflow-monitoring data through MCP and create scheduled digests, but scheduled checks are not sufficient for real-time incident response. If a workflow breaks or error and handoff rates spike between scheduled runs, the team may not learn about it until much later.

Customers need native, real-time or near-real-time alerts that can detect threshold breaches or sudden increases in errors and handoffs at the workflow or step level, then notify the team in Slack. Alerts should identify the affected workflow, step, metric, and change so teams can investigate and respond immediately.