VK Cloud has introduced a second generation of managed Kubernetes clusters, shifting the management of control planes, worker nodes, and system components to the platform. Unlike the first generation based on OpenStack Magnum, this model isolates infrastructure in dedicated service projects.
Users can manage clusters via the web interface or Terraform using the vkcs_kubernetes_cluster_v2 resource. The platform supports standard and fault-tolerant topologies, with master nodes automatically scaling vertically based on CPU and RAM usage thresholds.
The Cluster Autoscaler is available for worker node groups, supporting up to 500 nodes per group. When enabled, it monitors pod resource requests to trigger scaling. According to VK Cloud, this mechanism can reduce infrastructure costs by up to 60 percent.
The platform provides a catalog of add-ons, including Argo CD, ingress-nginx, and cert-manager, which are deployed as Kubernetes resources. Unlike eksctl, which relies on CloudFormation stacks, the VK Cloud orchestrator continuously maintains the desired state of the cluster.
