How can you implement a scalable solution for global Prometheus querying across multiple geographically distributed Google Kubernetes Engine (GKE) clusters while minimizing management overhead, given that your current Prometheus-based monitoring setup lacks global visibility?