The OpenShift autoscaler does not trigger a scale-up for a MachineAutoscaler with "minReplicas: 0" for Pods that define ephemeral-storage requests.

The OpenShift autoscaler does not trigger a scale-up for a MachineAutoscaler with "minReplicas: 0" for Pods that define ephemeral-storage requests.

So given the following MachineAutoscaler / MachineSet combination:

```
apiVersion: "autoscaling.openshift.io/v1beta1"
kind: "MachineAutoscaler"
metadata:
  name: sbb-worker-highmem-3-0
  namespace: "openshift-machine-api"
spec:
  minReplicas: 0
  maxReplicas: 3
  scaleTargetRef:
    apiVersion: machine.openshift.io/v1beta1
    kind: MachineSet
    name: sbb-worker-highmem-3-0
```

```
apiVersion: machine.openshift.io/v1beta1
kind: MachineSet
metadata:
  labels:
    machine.openshift.io/cluster-api-cluster: eap01p-2jsxv
  name: sbb-worker-highmem-3-0
  namespace: openshift-machine-api
spec:
  selector:
    matchLabels:
      machine.openshift.io/cluster-api-cluster: eap01p-2jsxv
      machine.openshift.io/cluster-api-machineset: sbb-worker-highmem-3-0
  template:
    metadata:
      labels:
        machine.openshift.io/cluster-api-cluster: eap01p-2jsxv
        machine.openshift.io/cluster-api-machine-role: highmem
        machine.openshift.io/cluster-api-machine-type: highmem
        machine.openshift.io/cluster-api-machineset: sbb-worker-highmem-3-0
    spec:
      metadata:
        labels:
          node-role.kubernetes.io/app: ""
          node-role.kubernetes.io/highmem: ""
      providerSpec:
        value:
          apiVersion: machine.openshift.io/v1beta1
          credentialsSecret:
            name: azure-cloud-credentials
            namespace: openshift-machine-api
          image:
            offer: ""
            publisher: ""
            resourceID: /resourceGroups/eap01p-2jsxv-rg/providers/Microsoft.Compute/images/eap01p-2jsxv
            sku: ""
            version: ""
          kind: AzureMachineProviderSpec
          location: westeurope
          managedIdentity: eap01p-2jsxv-identity
          networkResourceGroup: sbb-prod-eap-ocp-network
          osDisk:
            cachingType: ReadOnly
            diskSettings:
              ephemeralStorageLocation: Local
            diskSizeGB: 400
            managedDisk:
              storageAccountType: Standard_LRS
            osType: Linux
          publicIP: false
          publicLoadBalancer: eap01p-2jsxv
          resourceGroup: eap01p-2jsxv-rg
          securityProfile:
            encryptionAtHost: true
          subnet: sbb-prod-eap-ocp-private-sub
          userDataSecret:
            name: worker-user-data
          vmSize: Standard_E16as_v4
          vnet: sbb-prod-eap-ocp-vnet
          zone: "3"
      taints:
      - effect: NoSchedule
        key: node-role.kubernetes.io/highmem
```

Trying to start a pod that specifies ephemeral-storage requests...

```
   ...
    resources:
      requests:
        memory: "64Mi"
        ephemeral-storage: "256Mi"
        cpu: "250m"```
  ...
```

...fails with the following message from the autoscaler:

```
pod didn't trigger scale-up: 3 Insufficient "ephemeral-storage"
```

The problem itself is nothing special, the same holds true for CPU and memory when scaling from zero. Nevertheless, the current autoscaler implementation supports the "machine.openshift.io/memoryMb" and "machine.openshift.io/vCPU" annotations that help the autoscaler to determine which MachineSet to scale. But no such annotation seems to exist for ephemeral storage.

What options are there to trigger the scale-up of an autoscaled MachineSet with minReplicas 0, that currently has 0 active replicas, for a pod that specifies ephemeral-storage requests? The expected behavior would be that an annotation can be added to the MachineSet to infor the autoscaler about the amount of ephemeral storage on nodes created from the MachineSet. Always keeping a node running in the MachineSet is not an option.