Skip to content
VirtConsole Docs

Kubernetes clusters

Managed Kubernetes gives you a full cluster without installing anything by hand. The panel creates the virtual machines, installs Kubernetes on them, wires up networking, and puts the API server behind a managed load balancer. Provisioning typically takes 5 to 12 minutes.

For each cluster the panel creates one or three control plane nodes, your worker nodes in node pools, a load balancer fronting the API server, a per-cluster security group, and a kubeconfig file you download to use kubectl.

Before you begin

Section titled “Before you begin”
  • A location with Kubernetes enabled by your provider. The create wizard only lists enabled regions.
  • A VPC in that region, with a NAT gateway attached (workers need outbound internet to pull container images) and at least one private subnet. See VPC networks.
  • Enough credit balance if your provider bills hourly; the wizard shows the estimated hourly cost.

Where to find it

Section titled “Where to find it”

In the user panel sidebar go to Services > Clusters. The page still carries a Kubernetes > Clusters breadcrumb once open.

Kubernetes clusters list

Create a cluster

Section titled “Create a cluster”

Click Create cluster. A four-step horizontal wizard opens: Cluster, Networking, Node pool, Review.

Create Cluster wizard

  1. Cluster (identity and version):

    • Cluster name and Slug (used in node hostnames; lowercase letters, numbers and hyphens only). An optional Description.

    • Region: only regions with VPC, load balancer and Kubernetes enabled are listed. Locked regions your account cannot use yet appear below the picker with a request-access card.

      Region and VPC selected in the Create Cluster wizard

    • Kubernetes version: pick from your provider’s curated list; the newest is marked Recommended.

    • Control plane: 1 node (Single) (lower cost, no redundancy) or 3 nodes (HA) (survives a single-node failure).

    • Control plane plan and Control plane load balancer plan: the sizing tiers for the control plane VMs and the API server’s load balancer.

  2. Networking (VPC and subnets):

    • VPC: only VPCs in the chosen region are listed. Once picked, the page shows whether the VPC has a NAT gateway; workers need one for outbound internet access.
    • Control plane subnet (private subnets only) and Worker subnet (defaults to the control plane subnet; pick a different one to isolate workers).
    • Endpoint mode: Private (apiserver reachable only inside the VPC) or Public & Private (internal plus public load balancer endpoint). See Access and security for what each mode means.
    • Pod CIDR (default 10.244.0.0/16) and Service CIDR (default 10.96.0.0/12).
    • Pod Security Admission default: Privileged, Baseline (recommended) or Restricted.

  3. Node pool (the cluster’s first pool; add more from the Node pools tab once the cluster exists):

    • Worker plan.
    • Initial worker count, Min size and Max size.
    • Enable cluster autoscaler toggle; Min/Max size only apply while it is on.

  4. Review: check the summary rows and the estimated hourly cost (when your provider bills hourly), then click Create cluster.

Watch progress on the cluster’s Tasks tab; every operation is listed with its logs.

Use the cluster

Section titled “Use the cluster”

  1. Open the cluster and click Download kubeconfig.

  2. Point kubectl at it:

    Terminal window
    export KUBECONFIG=~/Downloads/kubeconfig-mycluster.yaml
    kubectl get nodes

A Service of type=LoadBalancer automatically provisions a managed load balancer and gets its public IP as EXTERNAL-IP. The annotations that control it are documented on Load balancer Services.

The cluster page header shows status, region, version, node counts and endpoints. The tabs, in order, are:

  • Overview: API endpoints, compute plans, networking and ownership at a glance.
  • Node pools: add and edit pools, set per-pool plans, sizes, autoscaling bounds, labels and taints. See Node pools.
  • Nodes: every VM in the cluster, with CPU and memory gauges and the pods on each node.
  • Tasks: every operation against the cluster, with logs. First stop when a deploy or upgrade fails.
  • Upgrades: rolling upgrades for the control plane and workers. See Upgrades and rotation.
  • Autoscaler: the copy-paste manifest that enables automatic worker scaling. See Autoscaling.
  • Security: the per-cluster security group in three scopes (load balancer, control plane, workers). See Access and security.
  • SSL & domains: bind a TLS certificate to the API load balancer for a custom domain. See Access and security.

Common problems

Section titled “Common problems”
  • The cluster is stuck in Provisioning. Open the Tasks tab and read the failing task’s log. Check the VPC’s NAT gateway is healthy; workers cannot finish bootstrap without outbound access.
  • kubectl gets connection refused. Confirm the cluster is Running, then work through Troubleshooting.
Section titled “Related”