Not detecting readiness of CNPG Cluster Resource

[raweber42]

What happened?

I'm using function-auto-ready with function-kcl to create CloudNativePG (CNPG) Cluster resources. The CNPG Cluster resources have the correct status.conditions with type: Ready, status: "True" when they become healthy, but function-auto-ready is not detecting them as ready.

Currently, I'm forced to hardcode resources as ready using the krm.kcl.dev/ready = "True" annotation from function-kcl, but this defeats the purpose - I want resources to only be marked as ready when they actually report ready status conditions.

Expected behavior: function-auto-ready should observe the CNPG Cluster resource's status and mark the composed resource as ready when the Cluster reports type: Ready, status: "True" in its status.conditions.

Actual behavior: The composed resource "postgres-cluster" continuously shows as "not yet ready" even when the underlying CNPG Cluster is in a healthy state with Ready=True condition.

Error message seen:

Normal  ComposeResources  63s (x1435 over 20h)    defined/compositeresourcedefinition.apiextensions.crossplane.io  Composed resource "postgres-cluster" is not yet ready

How can we reproduce it?

1. Create a Composition using Pipeline mode with function-kcl and function-auto-ready:

apiVersion: apiextensions.crossplane.io/v1
kind: Composition
metadata:
  name: postgres-abstraction
spec:
  compositeTypeRef:
    apiVersion: abc.io/v1alpha1
    kind: XPostgres
  mode: Pipeline
  pipeline:
  - step: create-resources
    functionRef:
      name: function-kcl
    input:
      apiVersion: krm.kcl.dev/v1alpha1
      kind: KCLInput
      spec:
        source: |
          oxr = option("params").oxr
          
          cluster = {
              apiVersion = "postgresql.cnpg.io/v1"
              kind = "Cluster"
              metadata = {
                  name = "${oxr.metadata.name}-pg"
                  namespace = "database"
                  annotations = {
                      "krm.kcl.dev/composition-resource-name" = "postgres-cluster"
                      "crossplane.io/external-name" = "${oxr.metadata.name}-pg"
                  }
              }
              spec = {
                  instances = 1
                  imageName = "ghcr.io/cloudnative-pg/postgresql:17.7-system-trixie"
                  # ... rest of spec
              }
          }
          
          items = [cluster]
  
  - step: automatically-detect-ready-composed-resources
    functionRef:
      name: function-auto-ready

2. Create an instance of the XR
3. Wait for the CNPG Cluster to become ready (it will have status.conditions with type: Ready, status: "True")
4. Observe that function-auto-ready still reports the composed resource as "not yet ready"

Actual CNPG Cluster status (showing it IS ready):

status:
  conditions:
  - lastTransitionTime: "2026-01-14T09:29:02Z"
    message: Cluster is Ready
    reason: ClusterIsReady
    status: "True"
    type: Ready
  phase: Cluster in healthy state
  readyInstances: 1

What environment did it happen in?

• Function version: v0.6.0 (also tested with v0.5.0)
• Crossplane version: Latest
• Kubernetes distribution: K3s (40+ on-premise clusters)
• Other functions in pipeline: function-kcl, function-status-transformer
• CRD: CloudNativePG Cluster (postgresql.cnpg.io/v1)

Additional Context

I'm not sure if this is:

1. A compatibility issue between function-kcl and function-auto-ready
2. function-auto-ready not properly observing custom resources (CNPG Clusters) in Pipeline mode
3. Missing configuration or annotation needed for function-auto-ready to observe these resources

I tried:

• Adding crossplane.io/external-name annotation
• Using krm.kcl.dev/composition-resource-name annotation
• Checking that CNPG Clusters expose the standard Ready condition

The only workaround is hardcoding "krm.kcl.dev/ready" = "True" in the KCL function output, but this marks resources as ready immediately upon creation rather than waiting for actual readiness.

[bobh66]

Have you enabled debug in the function and checked the logs? There are a lot use useful debug messages in the code.