25% of the crio/kubelet targets down alert shown in kube-system namespace for windows node

Issue

• After adding windows nodes in the cluster, below alert is constantly being fired:

25% of the crio/kubelet targets in kube-system namespace are down

• Prometheus shows target down with the below error for all the windows nodes endpoints:

http://nodeIP:9537/metrics  DOWN    
endpoint="crio"instance="nodeIP:9537"job="crio"namespace="kube-system"node="node-name"service="kubelet"
56.719s 1.485ms Get "http://nodeIP:9537/metrics": dial tcp nodeIP:9537: connect: connection refused