Managament cluster Grafana should use the Thanos datasource instead of the default Prometheus one provisioned by the Rancher monitoring stack