Advanced AI Prompt for Hybrid Cloud Zero-Downtime Synchronization

Act as a Hybrid Cloud Architect. Design a highly complex GitHub Actions workflow orchestrating a massive zero-downtime deployment spanning disparate infrastructure. The pipeline must trigger simultaneous code deployments to both a bare-metal [ON_PREM_CLUSTER] (via self-hosted runners on VPN) and a managed AWS/GCP [CLOUD_CLUSTER] (via GitHub-hosted runners). You must utilize [SYNC_TOOL] (e.g., ArgoCD multi-cluster) to verify both physical environments report identical ReplicaSet states. Before instructing the global load balancer to split user traffic, the workflow must execute a highly specific [DATA_CONSISTENCY_CHECK] script validating that the on-prem database and the cloud database are perfectly synchronized. Document the precise failover logic: if the cloud cluster fails the health check, how does the pipeline instantly route 100% of traffic back to the on-prem firewall?