Deploying CRD using terraform causes the error Forbidden attribute key in "manifest" value

[slok]

Deploying CRD using terraform causes the error Forbidden attribute key in "manifest" value

[slok]

Hi @Lexmark-jrichey

The CRDs are autogenerated with official Kubernetes tooling (https://github.com/kubernetes/code-generator). As you may see in these examples of official and very common deployed stacks, also have it:

• https://github.com/aws/eks-charts/blob/a78f6372c9739ed4a2472ee2eeba4cb8810a0a93/stable/aws-load-balancer-controller/crds/crds.yaml#L469
• https://github.com/prometheus-operator/kube-prometheus/blob/9b532dac0c8b00e984d2f77b2f2f0ef0dae6e751/manifests/setup/0prometheusCustomResourceDefinition.yaml#L8128
• https://github.com/kubernetes-sigs/application/blob/c8e2959e57a02b3877b394984a288f9178977d8b/deploy/kube-app-manager-aio.yaml#L537
• https://github.com/aws/amazon-vpc-cni-k8s/blob/235567459fa3b4d9e051eef418dbe349994c1a94/config/master/calico-operator.yaml#L174
• https://github.com/kubernetes-csi/external-snapshotter/blob/d74ae665fde0502c32135a879ce4549d029be1ee/client/config/crd/snapshot.storage.k8s.io_volumesnapshots.yaml#L228

I'm not saying that is ok, however, ignoring status it's a minor thing, that doesn't hurt anyone and reduces complexity on all sides, even, other official tools like kubectl/apiserver are already doing it.

So... IMHO, this is a problem on terraform side: hashicorp/terraform-provider-kubernetes#1428