Mastering Kubernetes Resources with Java: EndpointSlice, PVC, PV, NetworkPolicy & More

The article provides a comprehensive tutorial for managing Kubernetes resources using the Fabric8 Java client. It covers the full lifecycle—loading from YAML files, retrieving from the API server, creating new objects, applying changes with server‑side apply, listing resources (by namespace, across all namespaces, or by label), watching for events, and deleting objects.

EndpointSlice

Load from YAML:

EndpointSlice es = client.discovery().v1().endpointSlices()
    .load(getClass().getResourceAsStream("/endpointslice.yml")).item();

Get from API:

EndpointSlice esFromServer = client.discovery().v1().endpointSlices()
    .inNamespace("default").withName("es1").get();

Create:

EndpointSlice esToCreate = new EndpointSliceBuilder()
    .withNewMetadata().withName(name)
    .addToLabels("kubernetes.io/service-name", "example")
    .endMetadata()
    .withAddressType("IPv4")
    .addNewPort().withName("http").withPort(80).endPort()
    .addNewEndpoint()
        .withAddresses("10.1.2.3")
        .withNewConditions().withReady(true).endConditions()
        .withHostname("pod-1")
        .addToTopology("kubernetes.io/hostname", "node-1")
        .addToTopology("topology.kubernetes.io/zone", "us-west2-a")
    .endEndpoint()
    .build();
esToCreate = client.discovery().v1().endpointSlices().inNamespace("ns1").resource(esToCreate).create();

Apply to cluster:

client.discovery().v1().endpointSlices().inNamespace("ns1").resource(endpointSlice).serverSideApply();

List in a namespace, across all namespaces, or by label:

EndpointSliceList esList = client.discovery().v1().endpointSlices().inNamespace("default").list();
EndpointSliceList esListAll = client.discovery().v1().endpointSlices().inAnyNamespace().list();
EndpointSliceList esByLabel = client.discovery().v1().endpointSlices().inNamespace("default").withLabel("foo", "bar").list();

Delete:

client.discovery().v1().endpointSlices().inNamespace("default").withName("test-es").delete();

Watch:

client.discovery().v1().endpointSlices().inNamespace("default").watch(new Watcher<>() {
    @Override
    public void eventReceived(Action action, EndpointSlice resource) {
        // handle event based on action
    }
    @Override
    public void onClose(WatcherException cause) {
        // handle close
    }
});

PersistentVolumeClaim (PVC)

Load from YAML:

PersistentVolumeClaim pvc = client.persistentVolumeClaims().load(new FileInputStream("pvc.yaml")).item();

Get from API:

PersistentVolumeClaim pvc = client.persistentVolumeClaims().inNamespace("default").withName("test-pv-claim").get();

Create:

PersistentVolumeClaim persistentVolumeClaim = new PersistentVolumeClaimBuilder()
    .withNewMetadata().withName("test-pv-claim").endMetadata()
    .withNewSpec()
        .withStorageClassName("my-local-storage")
        .withAccessModes("ReadWriteOnce")
        .withNewResources().addToRequests("storage", new Quantity("500Gi")).endResources()
    .endSpec()
    .build();
client.persistentVolumeClaims().inNamespace("default").resource(persistentVolumeClaim).create();

Apply to cluster:

client.persistentVolumeClaims().inNamespace("default").resource(pvcToCreate).serverSideApply();

List and delete follow the same pattern as EndpointSlice, using .list() and .delete() methods.

PersistentVolume (PV)

Load from YAML, get, create, apply, list, and delete are demonstrated with analogous code snippets, including node affinity configuration in the creation example.

NetworkPolicy

Examples show loading from YAML, retrieving, building a policy with pod selectors, ingress rules, ports, applying, listing, and deleting.

PodDisruptionBudget (PDB)

Loading, retrieving, creating (with maxUnavailable and selector), applying, listing, and deleting are covered.

Access Review Resources

SelfSubjectAccessReview, SubjectAccessReview, and LocalSubjectAccessReview are created with builders, submitted via .create(), and the result status is printed.

RBAC Resources

ClusterRole, ClusterRoleBinding, Role, and RoleBinding are each shown with loading from YAML, creating via builders, retrieving, listing (including label filtering), and deleting. The ClusterRoleBinding example also demonstrates constructing Subject and RoleRef objects before creation.

All code snippets are ready to be copied into a Java project that uses the Fabric8 Kubernetes client, providing a practical reference for developers working with Kubernetes programmatically.