Never ending reconciliation on a Deployment due to env expected to be '[]' vs null

[Manny2014]

Description

I've created a simple ResourceGraph (below) and not thinking much about i set the spec.template.containers[0].env to equal to '[]' in the definition. The controller never reconciled b/c it kept finding and issue with the value of 'env' (desired '[]' vs. observed null).

apiVersion: kro.run/v1alpha1
kind: ResourceGraphDefinition
metadata:
...
spec:
  schema:
    apiVersion: v1alpha1
    kind: SimpleApp
    spec:
      # Basic types
      name: string | required=true description="Application name"
      replicas: integer | default=1 # minimum=1 maximum=100
      namespace: string | required=true
      image: string | required=true
      port: integer | default=8080
    status:
      # Status fields with auto-inferred types
      deploymentConditions: ${deployment.status.conditions}
      availableReplicas: ${deployment.status.availableReplicas}
      workloadKsa: ${ksa.metadata.name}
  resources:
  - id: ksa
    template:
      apiVersion: v1
      kind: ServiceAccount
      metadata:
        name: sv-${schema.spec.name}-ksa
        namespace: ${schema.spec.namespace}
 
  - id: deployment
    readyWhen:
      - ${deployment.spec.replicas == deployment.status.availableReplicas}
    template:
      apiVersion: apps/v1
      kind: Deployment
      metadata:
        name: ${schema.spec.name}
        namespace: ${schema.spec.namespace}
        labels:
          app.kubernetes.io/name: ${schema.spec.name}
      spec:
        replicas: ${schema.spec.replicas}
        selector:
          matchLabels:
            app.kubernetes.io/name: ${schema.spec.name}
            app: ${schema.spec.name}
        template:
          metadata:
            labels:
              app.kubernetes.io/name: ${schema.spec.name}
              app: ${schema.spec.name}
          spec:
            serviceAccountName: ${ksa.metadata.name}
            containers:
            - name: main
              image: ${schema.spec.image}
              imagePullPolicy: Always
              ports:
              - containerPort: ${schema.spec.port}
              resources:
                requests:
                  memory: "64Mi"
                  cpu: "250m"
                limits:
                  memory: "1Gi"
                  cpu: "1"
              env: []
            restartPolicy: Always

Observed Behavior:
The controller never reconciled and the Instance never made it past IN_PROGRESS although all objects were created successfully.

Watching the object:

NAME                         STATE         SYNCED   AGE
simpleapp.kro.run/simple-1   IN_PROGRESS   False    2m58s

NAME                       READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/simple-1   1/1     1            1           2m55s
[2]  + done       ~/local/google-cloud-sdk/bin/kubectl.1.30 get simpleapp,deploy -n rg-instance
[2] 30970
NAME                         STATE         SYNCED   AGE
simpleapp.kro.run/simple-1   IN_PROGRESS   False    3m1s

NAME                       READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/simple-1   1/1     1            1           2m58s

After removing 'env: []' from ResourceGraph the Instance reconciled in seconds:

NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/simple-1 1/1 1 1 3s
[2] + done ~/local/google-cloud-sdk/bin/kubectl.1.30 get simpleapp,deploy -n rg-instance
[2] 34541
NAME STATE SYNCED AGE
simpleapp.kro.run/simple-1 ACTIVE True 9s

NAME READY UP-TO-DATE AVAILABLE AGE
deployment.apps/simple-1 1/1 1 1 6s

Expected Behavior:
I did not expect that such a small thing would put the controller in this state.

Reproduction Steps (Please include ResourceGraphDefinition and Instances files):

Versions:

• kro version: 0.2.1
• Kubernetes Version (kubectl version):
  Client Version: v1.30.8
  Kustomize Version: v5.0.4-0.20230601165947-6ce0bf390ce3
  Server Version: v1.30.6+k3s1

Involved Controllers:

• Controller URLs and Versions (if applicable):

Error Logs (if applicable)**:

2025-02-11T03:36:58.790Z	DEBUG	controller.simpleapps	Found deltas for resource	{"namespace": "rg-instances", "name": "simple-1", "resourceID": "deployment", "delta": [{"path":"spec.template.spec.containers[0].env","desired":[],"observed":null}]}
2025-02-11T03:36:58.815Z	DEBUG	dynamic-controller	Finished syncing resourcegraphdefinition instance request	{"gvr": "kro.run/v1alpha1/simpleapps", "namespacedKey": "rg-instances/simple-1", "duration": "45.484577ms"}

• 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?

No response

[barney-s]

Fix proposed