Create and manage Service

A Service selects a set of Pods by labels and provides them with a stable access address. When workloads are scaled or Pods change, clients can still access Pods that meet the selection criteria through the Service. To configure HTTP or HTTPS routing with Ingress, you must first create a Service as the backend.

​

Create Service

1. Log in to the GPU Cloud console and go to the Clusters page.
2. Click the target cluster name to open the cluster workspace.
3. In the left navigation pane, go to the Network page and switch to the Services tab.
4. Click Create.
5. Configure Service.

   Parameter	Description
   Name	Set the Service name. The name should reflect the application or workload being accessed.
   Namespace	Select the namespace to which the Service belongs. When selecting Pods by labels, the Service and Pods must be in the same namespace.
   Labels	Optional. Add labels to the Service resource to identify or filter Services. The labels here do not determine which Pods traffic is forwarded to.
   Spec (YAML)	Enter the Kubernetes spec for the Service. Enter only the content under spec; you do not need to enter apiVersion, kind, or metadata. Here, selector is used to select backend Pods, and ports is used to configure the mapping between Service ports and Pod ports.

   The following example creates a Service of type ClusterIP, forwarding port 80 of the Service to port 8080 of Pods labeled app: example-app. ClusterIP is suitable for accessing workloads within the cluster and can also serve as a backend Service for Ingress.
   type: ClusterIP
   selector:
     app: example-app
   ports:
     - name: http
       protocol: TCP
       port: 80
       targetPort: 8080
   
   Modify the example based on the actual labels and listening ports of the workload:

   Field	Description
   type	The Service type. This example uses ClusterIP, which provides an access endpoint only within the cluster.
   selector	Selects the backend Pods. The key-value pairs must match the labels of the target Pods; otherwise, the Service cannot forward requests to the target Pods.
   port	The port exposed by the Service to clients.
   targetPort	The port on which the application in the backend Pod actually listens.

6. Click Create, then verify that the Service has been created.

​

View Service

1. On the Services tab, filter by namespace or use the Search box to locate the target Service.
2. In the Actions column of the target Service, click View.
3. Check the Service type, label selector, and port mappings to ensure that they point to the expected Pods and application ports.

If the Service has been created but the workload cannot be accessed, first check whether selector matches the Pod labels and whether targetPort is the port actually listened on by the application.

​

Delete Service

1. On the Services tab, locate the target Service.
2. In the Actions column for the target Service, click Delete.
3. Confirm the deletion.

When HTTP or HTTPS access needs to be provided through a domain name or path, see Expose Cluster Services Using Ingress.