1. For the Linux nodes, the container runtime is CRI-O and the port 9537 has a crio process listening on it.While, windows nodes doesn't have CRIO container runtime.

2. Prometheus is trying to connect to /metrics endpoint on the windows nodes on port 9537 which actually does not have any process listening on it. 

3. TargetDown is alerting crio job since it cannot reach the endpoint http://windows-node-ip:9537/metrics.

    1. Install 4.13 cluster with windows operator
    2. In the Prometheus UI, go to > Status > Targets to know which targets are down.   
    

    It gives the alert for targetDown

    It should not give any such alert.