How to Configure Ray to Switch to Different Node Group if first one was not available

We are using KubeRay and we have two NodeGroups in the Ray Cluster (Pods with different RAM and CPU limits to utilize all resources on two diffrent instances types). One issue that we are facing is that Ray Autoscaler choose one of Node Groups and if instance from that node group are not available, it does not try the other node group and waits fro ever for pod to be created. Is there a way to set a timeout, such that Ray Autoscaler tries next node group?