Take a single node cluster, the HCO will see the control plane is not highly available and set the kubevirt object with "infra: replicas: 1" as per [1]. That results in deployments with single replica for virt-api, 1 virt-controller etc as per [2].

Now look at the monitoring logic to fire alerts in [3], it doesn't look at the control pane availability of the cluster, it looks at the number of worker nodes and decides to fire the alerts if each pod has 1 replica only and the cluster has 2+ worker nodes.

This is conflicting logic, one side decides to scale or not based on X (control plane high availability), and the other complains or not based on Y (number of nodes).

Take a SNO cluster and add a single worker and this discrepancy will show up, as the control plane is not highly available but there are 2+ workers.

Please investigate what is the correct action here, if these alerts should not fire or if the components should start scaling.

[1] https://github.com/kubevirt/hyperconverged-cluster-operator/blob/cf24cf1dcc5fadf5ee53ed0fee7c005e42d6e66c/controllers/operands/kubevirt.go#L809
[2] https://github.com/kubevirt/kubevirt/blob/133430eab48bb567535f8556356e3b071b275388/pkg/virt-operator/resource/apply/apps.go#L56
[3] https://github.com/kubevirt/kubevirt/blob/15090de920b2345df8f56eb70a3d6ecc64d60992/pkg/monitoring/rules/alerts/virt-controller.go#L64

4.16.4

Always

1. Install SNO cluster
2. Add one extra worker node.

Alerts like LowVirtAPICount and LowVirtControllersCount  firing 

Either starts scaling or don't complain that its not scaling.