Kubernetes

Managed Kubernetes provisions a cluster in the region you choose and lets you run your workloads with kubectl. Clusters are scoped to your organization, and you connect to them with a downloadable kubeconfig.

Create a cluster

In the console, go to Kubernetes → Create cluster and fill in:

• Name — lowercase letters, digits and hyphens. It must start and end with an alphanumeric character and be at most 63 characters (an RFC 1123 label), for example prod-eu.
• Region — pick one from the available list.
• Version — 1.30 (default), 1.29 or 1.28.

The cluster list shows each cluster's name, version, node count, status and region.

When you create a cluster it starts in the provisioning state. Once it finishes, the status becomes running.

Connect with kubectl

After the cluster has finished provisioning, download its kubeconfig and point kubectl at it:

export KUBECONFIG=~/Downloads/<name>.kubeconfig.yaml
kubectl get nodes
kubectl get pods -A

The kubeconfig is also available programmatically over the signed gRPC API (nexlayer.k8s.v1.K8sService/GetClustersKubeconfig), returned as YAML. The kubeconfig becomes available only after the cluster finishes provisioning — while the cluster is still being created the call returns kubeconfig not yet available.

WARNING

The kubeconfig grants full access to your cluster. Keep it secret — treat it like a password and never commit it to source control.

Node pools and scaling

Each cluster has a node pool. To scale it, set the desired node count in the console. The value is clamped to the platform maximum, and the pool's status becomes scaling while the change is applied. Programmatically this is K8sService/CreatePoolsScale with { "node_count": 3 }.

Delete a cluster

Deleting a cluster (from the console, or K8sService/DeleteClusters over the API) sets its status to deleting and tears down its nodes.

WARNING

Deleting a cluster removes its nodes and the workloads running on them. This cannot be undone.

Cluster statuses

• provisioning — the cluster is being created.
• running — the cluster is ready to use.
• scaling — the node pool is changing size.
• deleting — the cluster is being removed.

API reference

The management API is gRPC: service nexlayer.k8s.v1.K8sService, called with a signed Aral Cloud client and a Bearer token — the same transport the console uses.

RPC	Request
CreateClusters	{ name, region, version }
ListClusters	—
GetClustersKubeconfig	{ id }
DeleteClusters	{ id }
CreatePoolsScale	{ id, node_count }

TIP

Prefer the terminal for everything else too? See the CLI guide.