When a regional incident strikes, every minute of downtime carries real customer impact — yet most organizations still rely on fragmented runbooks and manual coordination. This session presents EWOK, a production-grade, agent-based DR orchestration system built at Intuit to automate complex, stateful failovers across Active/Passive multi-region cloud architectures.

We'll walk through how EWOK replaces brittle DR scripts with declarative YAML workflows, an event-driven AWS State Machine, and a distributed Semaphore Service enforcing capacity-aware ordering across up to 100 concurrent multi-workload failovers — coordinating Kubernetes scaling, Aurora Global Database promotions, Route53 DNS updates, and Redis failovers while preventing GitOps configuration drift during live incidents.

Results from Intuit's critical financial platforms: MTTR reduced from 30+ minutes to under 5, zero idle hot-standby waste through just-in-time scaling, and 100% automated auditability. Attendees leave with a reusable architectural blueprint for production-hardened, platform-aware DR automation.