> ## Documentation Index
> Fetch the complete documentation index at: https://docs.rafftechnologies.com/llms.txt
> Use this file to discover all available pages before exploring further.

# raff_k8s_node_pool

> Manage additional Kubernetes node pools with Terraform — node count, autoscaling bounds, labels and taints.

<sub>Updated August 25, 2026</sub>

Manages an **additional node pool** on a Raff Kubernetes cluster. The cluster's first pool is defined on [`raff_k8s_cluster`](/reference/terraform/raff_k8s_cluster) (`default_pool`). All pools use the cluster's worker plan.

Scale-down drains nodes first, honouring PodDisruptionBudgets — a node that can't be emptied stops the operation. On a subscription cluster, growing a pool via Terraform charges the pro-rated difference to the saved payment method and raises the monthly price.

## Example

```hcl theme={null}
resource "raff_k8s_node_pool" "batch" {
  cluster_id = raff_k8s_cluster.prod.cluster_id
  name       = "batch"
  plan_id    = data.raff_k8s_node_plans.all.plans[0].id
  node_count = 2

  autoscale_enabled = true
  min_nodes         = 2
  max_nodes         = 6

  labels = jsonencode({ workload = "batch" })
  taints = jsonencode([{ key = "batch", value = "true", effect = "NoSchedule" }])
}
```

## Argument reference

| Argument            | Type   | Required? | Description                                                                       |
| ------------------- | ------ | --------- | --------------------------------------------------------------------------------- |
| `cluster_id`        | string | Yes       | Short cluster ID (`raff_k8s_cluster.x.cluster_id`)                                |
| `name`              | string | Yes       | Pool name, unique in the cluster                                                  |
| `plan_id`           | int    | Yes       | Worker node plan ID — must match the cluster's existing plan                      |
| `node_count`        | int    | Yes       | Node count (2–20)                                                                 |
| `autoscale_enabled` | bool   | No        | Enable autoscaling between `min_nodes` and `max_nodes`                            |
| `min_nodes`         | int    | No        | Autoscale lower bound                                                             |
| `max_nodes`         | int    | No        | Autoscale upper bound                                                             |
| `labels`            | string | No        | JSON object of Kubernetes labels applied to every node, e.g. `{"tier":"db"}`      |
| `taints`            | string | No        | JSON array of taints, e.g. `[{"key":"gpu","value":"true","effect":"NoSchedule"}]` |

## Attribute reference (computed)

| Attribute           | Description                                                                                                                          |
| ------------------- | ------------------------------------------------------------------------------------------------------------------------------------ |
| `id`                | Pool ID                                                                                                                              |
| `actual_node_count` | Nodes that currently exist — differs from `node_count` briefly while scaling, and when the autoscaler has added nodes above the base |

## Import

```bash theme={null}
terraform import raff_k8s_node_pool.batch <cluster-id>/<pool-id>
```