To allow the customer to omit certain files from the final must-gather, we give the option to filter by k8s resources: 

   # this omits all k8s Secrets
   - type: kubernetes
     kind: Secret
 
   # this omits all k8s ConfigMaps in the kube-system namespace
   - type: kubernetes
     kind: ConfigMap
     namespaces:
       - kube-system
 
   # this omits the let's encrypt custom resource "Certificate" with apiVersion "cert-manager.io/v1" in the given namespace
   - type: kubernetes
     apiVersion: cert-manager.io/v1
     kind: Certificate
     namespaces:
       - openshift-ingress-operator
 
   # this omits kubernetes resources in the below namespaces (not logs)
   - type: kubernetes
     namespaces:
       - openshift-node
       - openshift-kube-scheduler

In files where we can detect that, by knowing the schema in advance (or some fraction of it like the metadata), we can offer the ability to have semantic omission of the kubernetes resources. That allows for great flexibility, as demonstrated above, yet is quite complex to implement.

In scope for this story is:

• Namespace
• Kind
• API Version

All needs to be also used in conjunction, like the third example shows.

AC:

• Invocations of the CLI that have the kubernetes resource omission defined should omit the files from the final must-gather