`Kubectl.get(V1EndpointSlice.class)` failed due to `endpoints: null`

[fanyang01]

Describe the bug

It is possible that a service does not have any endpoint, for example, when the corresponding deployment was scaled to zero. In this case, the EndpointSlice API returns endpoints: null. However, it will be rejected by the following validation rule:

java/kubernetes/src/main/java/io/kubernetes/client/openapi/models/V1EndpointSlice.java

Lines 329 to 332 in cbd3d1b

	// ensure the json data is an array
	if (!jsonObj.get("endpoints").isJsonArray()) {
	throw new IllegalArgumentException(String.format("Expected the field `endpoints` to be an array in the JSON string but got `%s`", jsonObj.get("endpoints").toString()));
	}

Client Version
20.0.1

Kubernetes Version
1.29.0-eks-c417bb3

Java Version
Java 21

To Reproduce
Steps to reproduce the behavior:

• Create a service that cannot select any pod:

kubectl apply -f - <<EOF
apiVersion: v1
kind: Service
metadata:
  name: nginx-service
spec:
  selector:
    app.kubernetes.io/name: proxy
  ports:
  - name: name-of-service-port
    protocol: TCP
    port: 80
    targetPort: http-web-svc
EOF

• The endpoints field will be null:

k get endpointslice nginx-service-t786l -o yaml
addressType: IPv4
apiVersion: discovery.k8s.io/v1
endpoints: null
kind: EndpointSlice
...

Expected behavior

Kubectl.get(V1EndpointSlice.class) would not throw IllegalArgumentException.

Server (please complete the following information):

• OS: Bottlerocket
• Environment: container
• Cloud: AWS

[brendandburns]

Personally, I think this is an issue with how Go handles encoding empty arrays (https://medium.com/swlh/arrays-and-json-in-go-98540f2fa74e)

In practice, if you want to fix this, it needs to be in the upstream code generator, since this parsing is generated by that code generator.

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

[fanyang01]

@brendandburns Thanks for your comment. I am closing this issue. Here is my workaround:

            List<V1EndpointSlice> endpointSlices;
            try {
                endpointSlices = Kubectl.get(V1EndpointSlice.class)
                        .apiClient(client.apiClient())
                        .namespace(namespace)
                        .options(options)
                        .execute();
            } catch (IllegalArgumentException e) {
                String message = e.getMessage();
                if (message != null && message.contains("`endpoints`")) {
                    endpointSlices = null;
                } else {
                    throw e;
                }
            }