Skip to main content

Add node pools

A node pool is a group of worker nodes with identical configurations. A cluster can contain several node pools – this allows you to run workloads of different sizes within the same cluster without having to configure all nodes to be the same size.

Common reasons for having an extra swimming pool:

  • Memory-intensive workloads should run on larger nodes than stateless web services.
  • A workload should be scaled independently of the others.
  • You wish to switch to a different instance type without affecting the existing pool.

Prerequisites​

  • An existing cluster with the status Running
  • Sufficient resources in the subscribed plan

Add a node pool​

  1. Open the cluster in the Kubernetes overview.

  2. Switch to the Resources tab.

  3. Click on Add Node Pool +.

  4. Complete the form:

    FieldDescription
    Node pool nameName of the pool for identification in the overview and manifests.
    Machine Type (Instance)Machine category, such as Worker CPU or Premium Storage NVMe.
    Instance TypeInstance type C0 to C8 with vCPU, RAM, NVMe and price per node.
    NodesFixed number of nodes in the pool.
    Set node pool to autoscaleAlternatively, autoscaling with Minimum Nodes and Maximum Nodes.

  5. Confirm the creation.

The new pool initially appears in the list with the status Provisioning. As soon as all nodes are ready, the display changes to show the number of nodes running, for example 3/3 running.

Tipp

We recommend at least 2 nodes per pool to ensure your application remains accessible during upgrades and maintenance work.

Manage node pools​

On the Resources tab, you can see all node pools, along with their name, creation date, tags, number of nodes and autoscaling status.

  • Click the + icon next to the pool name to expand the list of nodes it contains.
  • Use the kebab menu (three dots) at the end of a row to access the pool’s actions.
  • Use the search field and the Filter button to narrow down the list by status, autoscaling and tags.

Managing individual nodes​

Once a pool has been expanded, the following are available per node via the Kebab menu:

ActionEffect
RecycleThe node is restarted. The pool replaces it with a fresh node.
DeleteThe node is removed.

In the process, nodes pass through the states Provisioning, Running and Deleting.

Hinweis

Move stateful workloads beforehand, or use a PodDisruptionBudget to ensure that a sufficient number of replicas remain accessible.

Targeted placement of workloads​

To ensure that an application runs on a specific pool, use a nodeSelector or affinity rules in your manifest:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: verarbeitung
spec:
  replicas: 2
  selector:
    matchLabels:
      app: verarbeitung
  template:
    metadata:
      labels:
        app: verarbeitung
    spec:
      nodeSelector:
        pool: speicher-pool
      containers:
        - name: worker
          image: registry.example.com/verarbeitung:1.0

You can determine the available labels for your nodes using:

kubectl get nodes --show-labels
Last updated on