HPA resource won't sync when using the averageValue field in custom metrics

[cliffpracht]

This may be an issue with the HPA controller and not ArgoCD, but it appears Argo has a problem with syncing an autoscaling policy that includes spec.metrics[0].object.target.averageValue field type. The HPA is created successfully but Argo believes the live manifest needs an additional field value and because git does not have said field defined in it, the resource never gets synced.

For instance a perfectly valid HPA manifest may look like this:

apiVersion: autoscaling/v2beta2
kind: HorizontalPodAutoscaler
metadata:
  name: hpa-name
spec:
  maxReplicas: 4
  metrics:
  - object:
      describedObject:
        apiVersion: apps/v1
        kind: Deployment
        name:  the-deployment-name
      metric:
        name: a-custom-metric
      target:
        averageValue: '1000'
        type: AverageValue
    type: Object
  minReplicas: 1
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: the-deployment-name

When argo applies this manifest, the HPA resource never reaches a synced/healthy state because of a perceived difference of value.

We were able to workaround this issue by adding the value field to the manifest in Git, but according to the HPA documentation this should not be needed.

Thoughts or suggestions on where to go from here?

On a side note, I don't fully understand how argo gets the applied manifest for HPA because it is different then if you were to look at the resource directly on the cluster (kubectl get hpa hpa-name -o yaml).