For the complete documentation index, see llms.txt. This page is also available as Markdown.

Clusters

Manage Kubernetes cluster lifecycle: create, list, inspect, delete, retrieve kubeconfig, check available versions, and list nodes.


Versions

List available Kubernetes versions for cluster creation.

ace k8s cluster versions

Example output:

Available Kubernetes Versions:
  v1.32.6
  v1.31.10
  v1.30.14

The CLI uses the display label (e.g. v1.32.6) and automatically resolves it to the full RKE2 version value (e.g. v1.32.6+rke2r1) at creation time.


Create

Create a new managed Kubernetes cluster.

ace k8s cluster create \
  --name <cluster-name> \
  --version <k8s-version> \
  --flavor <flavor-name-or-id> \
  --worker-name <pool-name> \
  [options]

Required Flags

Flag
Description

--name

Cluster name (lowercase alphanumeric with hyphens, max 63 chars)

--version

Kubernetes version (use ace k8s cluster versions to list)

--flavor

Flavor name or UUID for worker nodes (use ace flavor list)

--worker-name

Name for the initial worker node pool

Optional Flags

Flag
Default
Description

--worker-count

1

Number of worker nodes (1-8)

--volume-size

50

Volume size in GB per worker node (20-256)

--endpoint-access

Public

API endpoint access: Public, Private, or Public and Private

--network-isolation

Disabled

Network isolation: Enabled or Disabled

--nginx-ingress

Enabled

Nginx ingress controller: Enabled or Disabled

--nginx-default-backend

Enabled

Nginx default backend: Enabled or Disabled

--network-provider

Calico

Network provider (currently only Calico)

--snapshot-backup

No

Enable ETCD snapshot backup: Yes or No

--secrets-encryption

Disabled

Secrets encryption: Enabled or Disabled

--max-worker-nodes

8

Maximum number of worker nodes allowed

--security-group

Security group UUID to attach

--gpu

false

Mark as GPU flavor

--autoscale

false

Enable cluster autoscaling (can only be set at creation)

--autoscale-min

1

Minimum nodes for autoscaling

--autoscale-max

3

Maximum nodes for autoscaling

Examples

Minimal cluster:

Multi-node with custom volume size:

GPU cluster:

With autoscaling:

With security group and ETCD snapshots:

Note: Autoscaling can only be enabled at cluster creation time. It cannot be toggled after the cluster is created. When autoscaling is enabled, --autoscale-max must be less than or equal to --max-worker-nodes.


List

List all Kubernetes clusters in the current project and region.

Example output:

JSON output:


Get / Show

Get detailed information about a specific cluster. Accepts cluster name or ID.

Examples:

Example output:


Delete

Delete a Kubernetes cluster. Accepts cluster name or ID.

Examples:

Note: Deleting the last cluster in a project will tear down all K8s infrastructure for that project.


Kubeconfig

Generate a kubeconfig file for kubectl access to a cluster.

By default the YAML kubeconfig is written to stdout, ready to redirect into a file or pipe into kubectl.

Save to file:

Merge with existing config:

Output format

The --output / -o flag is honoured:

Format
Output

table (default)

Raw YAML kubeconfig on stdout — ready for redirection into ~/.kube/config.

yaml

Raw YAML kubeconfig (identical to table for this command).

json

A wrapper object: {"kubeconfig": "<yaml-as-string>"}. Useful for piping into jq or automation.

Verify access:


Nodes

List nodes in a Kubernetes cluster. When no --pool is specified, the CLI automatically queries all node pools and merges the results.

Flags

Flag
Description

--cluster

Alternative to positional argument for cluster name/ID

--pool

Filter nodes by a specific node pool name

Examples

List all nodes across all pools:

Filter by pool:

Example output:

The table view auto-hides columns that are uniformly empty across the row set (e.g. ID, IP, AGE are dropped when the backend hasn't populated them yet). Use -o json for the full per-node payload.

Full payload:


Global Flags

All K8s commands support these global flags:

Flag
Description

-o, --output

Output format: table (default) or json

-v, --verbose

Show full JSON response

--project

Override project ID

--region

Override region

Last updated