Node Pool Tab
The Node Pool tab manages worker capacity. From here you add, edit, resize, and delete node pools, view scaling policies, and act on individual worker nodes. The tab's content is available once the cluster is Running.
To open it, select a cluster and choose the Node Pool tab.
Each pool is shown as an expandable panel. Expand a pool to see its scaling configuration, plan and pricing details, and the worker-node table.
Add a Node Pool
Select Add Node Pool (requires update permission). The add-pool drawer opens with the same options as the create flow: pool name, plan/family, and pool type (Static or Autoscale) with its scaling policy.
A cluster can have up to 10 node pools. A newly added pool appears at the top of the list and takes a few minutes to deploy.
Pool Controls
Each pool panel header provides these controls (most require update permission and the pool to be Running):
| Control | When it appears | What it does |
|---|---|---|
| Refresh | Always | Reloads the pool's data. |
| Edit | Running and the pool is an autoscale pool | Edit the pool's autoscale range and scaling policy. |
| Resize | Running and the pool is not on a committed plan | Change the worker node count. |
| Committed Settings / Convert to Committed | Running/Warning, the plan supports committed, the pool is not autoscale, and not a private cluster | Manage committed billing for the pool. |
| Delete Pool | Always (delete permission) | Remove the entire pool. |
Edit a Pool
Available for autoscale pools. Open Edit to change the minimum/maximum worker count and the scaling policy (elastic, scheduled, or both). See Autoscaling for the policy fields.
Resize a Pool
Open Resize to set a new worker node count using the minus/plus controls.
- For an autoscale pool, the count is bounded by the pool's minimum and maximum, so the minus/plus buttons disable at those limits.
- If you increase an autoscale pool above its current size, a prompt asks whether to also raise the minimum desired count:
- Also update minimum desired nodes - prevents autoscaling from scaling back down to the old minimum.
- Keep minimum unchanged - nodes may be removed automatically later if utilization is low.
Confirm with Resize Pool.
Delete a Pool
Open Delete Pool. To confirm, type the exact node pool name in the confirmation field, then select Delete. If the pool is on a committed plan, a warning notes that the pool is reserved until the committed end date.
Pool Details
When expanded, each pool shows:
- Plan Name - the worker plan (or
<cluster name> (Private Cloud)for a private cluster). - Pool Type - Static or AutoScale.
- Pool State - the pool's current state, with a spinner during intermediate states (Creating, Scaling, Cooldown, Deploying).
- Specifications - vCPU, RAM, and storage for the plan (public clusters).
- Committed Plan / Hourly Plan - committed period and end date if committed, otherwise the hourly price.
- Committed Node Settings - end-of-term behavior (switch to hourly, auto-renew, or auto-deletion), shown only for committed pools.
Autoscaling Details
Autoscale pools also display:
- Minimum / Maximum Workers.
- Elastic Policy - the scale-up and scale-down rules (parameter, operator, and threshold), Watch Period, Cooldown (s), and Period duration (s). For a custom-parameter policy, a Get cURL link generates the command used to update the custom parameter.
- Scheduled Policy - Upscale Recurrence and desired nodes, and Downscale Recurrence and desired nodes.
See Autoscaling for what each field means and the custom-parameter cURL workflow.
Worker Nodes
The worker-node table lists every node in the pool:
| Column | Description |
|---|---|
| Name | The node's display name. Editable - select the edit icon, enter a new name, and save. The name must be 1–128 characters, start with a letter, end with a letter or digit, and contain only letters, digits, underscores, or hyphens. |
| Kubernetes Worker Name | The node's name inside Kubernetes (NA until assigned). |
| Private IP | The node's private IPv4 address, with a copy button. |
| Status | The node's state (Running, Powered off, Powering on, Deploying, etc.), with a spinner during transitions. |
| Created At | When the node was created. |
| Actions | Per-node actions (see below). |
While a pool is still deploying, the table shows Node pool is deploying. It will be available within 4–7 minutes.
Worker Node Actions
Open a node's three-dot Actions menu:
| Action | Available when | Effect |
|---|---|---|
| Start | The node is Powered off (and the pool is not in Cooldown) | Powers the node back on. |
| Power Off | The node is Running | Powers off the node. Pods on it are terminated and recreated on other running nodes. Persistent Volumes (PVCs) are not affected. |
| Reboot | The node is Running | Reboots the node. Its pods are terminated and recreated. PVCs are not affected. Disabled while Powered off, Powering on, Terminating, or Deploying. |
| Delete | The node is Running (delete permission) | Removes the node from the pool. Anything stored directly on the node is lost; PVCs are not affected. |
Each action opens a confirmation dialog naming the node. For a node in a committed pool, the Delete dialog warns that the node is reserved until the committed end date.
You cannot delete a node below the pool's minimum. The minimum is 2 for a compute (CPU) pool and 1 for a GPU pool. If you try, the portal shows: Cannot delete the node! The minimum node count for the pool is 2 (or 1). To shrink below the minimum, delete the pool instead.
Related Resources
| Resource | Use it for |
|---|---|
| Autoscaling | Elastic, custom, and scheduled scaling policies. |
| Committed Plans | Committed billing for pools. |
| Monitoring and Alerts | Per-node metrics and alerts. |
| Actions | Cluster-level lifecycle actions. |