Skip to content

Services

Use a service when you want a stable internal address for a deployment.

In RemoteGPU, services are normal Kubernetes services inside your namespace. They are your resources, whether you create them from the console or with kubectl.

Open in the console

TaskConsole page
Create or inspect servicesKubernetes / Services

Typical reasons to create a service

Use caseWhy a service helps
Internal app-to-app trafficGives other workloads a stable in-cluster target
Pod replacement and scalingKeeps the target stable while pods restart or scale
Future public publishingProvides the backend target for an ingress

Supported service profile

AreaSupportedNot supported
Service typeClusterIPNodePort
Public exposure modelIngress backed by a serviceexternalIPs
Ownership modelNamespace-scoped Kubernetes ServicePlatform-managed hidden service abstraction

Choose a workflow

WorkflowBest forWhat happens
ConsoleGuided setup for a deployment you already createdThe console reads the deployment selector and creates a normal ClusterIP service
kubectlFull native control over the service manifestYou define the selector and ports directly

Create a service

From the console:

  1. Open Kubernetes / Services
  2. Select a namespace
  3. Click Create service
  4. Choose the deployment you want to expose
  5. Provide the service details
Console fieldWhat it means
Service nameKubernetes service name in the namespace
Service portThe port clients use to reach the service
Target portThe container port or named target on the selected deployment

The console reads the deployment selector and creates a normal ClusterIP service for that deployment.

After that, it behaves like any other Kubernetes service in the namespace.

Console and kubectl work on the same service list

The Services page shows the live services that currently exist in the namespace.

That includes services created from the console and services created with kubectl.

If you update or delete a service with kubectl, the console reflects that live state after refresh.

Next step: publish the service

If the service should be reachable from the internet, create an ingress for it.

An ingress points to one or more existing service backends and gives you HTTP routing for public access.

RemoteGPU customer documentation