Advanced AI Prompt for Automated Rollback Logic with GitHub Actions

Act as a CI/CD Reliability Engineer. Develop a highly intelligent GitHub Actions workflow centered entirely around automated Mean Time To Recovery (MTTR). After a successful code deployment, the runner must *not* exit. Instead, it must initiate a 10-minute blocking 'monitoring window' loop. Within this loop, execute an API call to [MONITORING_TOOL] (e.g., Datadog, New Relic) to actively scan the [LOG_GROUP] for specific StackTrace or 500 error patterns. If the mathematical [ERROR_THRESHOLD] (e.g., > 50 errors/min) is breached during this window, the runner must aggressively catch the failure, execute the [ROLLBACK_COMMAND], post a high-priority incident payload to a Slack Webhook, and exit with status code 1. Detail how to persist the Git SHA of the previously stable build to ensure the rollback is instantaneous.