Unable to reconcile RGD if cluster contains unavailable APIService

[maxsxu]

Description

Observed Behavior:

The RGD inactive and has following error message

apiVersion: kro.run/v1alpha1
kind: ResourceGraphDefinition
status:
  conditions:
  - lastTransitionTime: "2025-03-25T15:09:52Z"
    message: Directed Acyclic Graph is synced
    reason: failed to retrieve Kubernetes namespaced resources: unable to retrieve the complete list of server APIs: custom.api.server/v1alpha1: stale GroupVersion discovery: custom.api.server/v1alpha1....
    status: "False"
    type: GraphVerified

Expected Behavior:

The RGD should be reconciled successfully as long as the RGD resources doesn't refer any unavailable APIService.

Reproduction Steps (Please include ResourceGraphDefinition and Instances files):

Using example in https://kro.run/docs/getting-started/deploy-a-resource-graph-definition

Versions:

• kro version: 0.2.2
• Kubernetes Version (kubectl version): v1.32.3

Involved Controllers:

• Controller URLs and Versions (if applicable):

Error Logs (if applicable)**:

2025-03-25T15:09:52.492Z    DEBUG    controller.resourceGraphDefinition    processing reconciliation error    {"resourcegraphdefinition": {"name":"my-application"}, "error": "failed to retrieve Kubernetes namespaced resources: unable to retrieve the complete list of server APIs: custom.api.server/v1alpha1: stale GroupVersion discovery: custom.api.server/v1alpha1....."}

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

Which option describes the most your issue?

ResourceGraphDefinition (Create, Update, Deletion)

[maxsxu]

After the unavailable APIServices back online, the RGD still unable to reconcile.

Resolved by recreating the RGD, but observed the CR/Application will show ReconciliationFailed status temporally at the beginning, and disappear after the Deployment ready.

apiVersion: kro.run/v1alpha1
kind: Application
status:
  conditions:
  - lastTransitionTime: "2025-03-25T15:41:38Z"
    message: 'failed to synchronize reconciling resource deployment: failed to evaluate
      dynamic variables: incomplete data: failed evaluating expression deployment.status.availableReplicas:
      no such key: availableReplicas'
    observedGeneration: 1
    reason: ReconciliationFailed
    status: "False"
    type: InstanceSynced
  state: ERROR

[newbhumika]

I can solve this issue . please assign it to me .

[vsavinkov]

I'm facing this too. In fact, KRO is stuck in case of stale GroupVersion unless you restart its pod

[joshuabaird]

I'm also hitting this. Restarting the kro controller seems to clear it up.

[github-actions]

This issue has been marked stale due to 30 days of inactivity. Maintainers use this label for hygiene. Comment or add a relevant label (keepalive) to keep it from becoming stale.