A university research portal serving 80 departments runs API Gateway, Lambda, and Aurora Serverless with 10x traffic during enrollment week, needing latency alerts, quota exhaustion warnings, and automatic remediation without paging on-call staff. Which health monitoring and automated response design provides early warning and self-healing capacity management?

The correct design works because CloudWatch alarms continuously evaluate workload health signals such as API latency and errors alongside Service Quotas usage metrics with thresholds near eighty percent, and alarm state changes emit events to Amazon EventBridge which routes them deterministically to AWS Systems Manager Automation runbooks that execute approved remediation such as requesting quota increases, throttling adjustments, or scaling actions without human intervention. The AWS Config rules approach fails by concept because Config evaluates configuration compliance periodically and notifies operators for manual action, so it lacks continuous health monitoring and automated response. The Application Insights plus scheduled scaling approach fails by concept because scheduled Auto Scaling handles predictable compute capacity but does not monitor Service Quotas exhaustion or trigger event-driven runbooks across diverse services. The Prometheus plus Grafana approach fails by concept because visualization and custom Lambda quota requests add operational overhead and do not provide native alarm-driven EventBridge to Systems Manager remediation with approvals and audit trails. Exam caveat: when the stem asks for early quota warning plus automatic runbook execution, look for CloudWatch alarms on quotas plus EventBridge to SSM Automation. Practical check: confirm Service Quotas CloudWatch metrics exist in the Region, set alarm thresholds below hard limits, and test the EventBridge to Automation execution role in a non-production surge rehearsal.