Release notes in Managed Kubernetes
2026
-
updated the list of supported Kubernetes versions;
-
updated the format for storing containerd configuration on worker nodes - now the configuration is stored in the
/etc/containerd/certs.d/docker.io/directory in thehosts.tomlfile . This allows you to add directories for your own registries and configure them separately, for example you can specify CA-certificate for self-signed TLS or disable certificate validation. For more information on customizing containerd registries via thehosts.tomlfile, see the containerd documentation on the GitHub site.For Managed Kubernetes clusters version 1.31 and higher, the new containerd configuration storage format is applied by default when the cluster is created. In existing clusters, the changes are applied automatically when reinstalling worker nodes or creating a new node group.
- added mks service events related to integration with Container Registry to audit logs;
- updated Karpenter provider - now Karpenter correctly calculates the amount of memory available after booting the operating system and selects flavors with enough resources to run pods;
- added
cni_typeparameter to selectel_mks_cluster_v1 resource - now network plugin type (CNI) can be selected when creating a cluster via Terraform.
- added mks service events to the audit logs;
- removed the limitation on network cards in clusters on dedicated servers - now you can use network cards with connection speeds of 10 Gbps and 25 Gbps;
- added support for ephemeralStorage when autoscaling with Cluster Autoscaler and Karpenter;
- added the ability to view in the control panel the user data you specified when creating a cluster on a cloud server or adding a new node group;
- have added the
pigzutility to the working node image, which speeds up the unpacking of container images through multi-threaded layer processing. The containerd runtime environment automatically starts using thepigzutility . You do not need to enable or configure anything else.
-
updated the autoscaling process with Karpenter:
- accelerated the scaling process through parallel creation of multiple NodeClaims;
- added UserData support to NodeClass;
- added PriorityClassName to the Helm chart on Karpenter installation.
2025
-
updated the Kubernetes version list: added version 1.34 and removed Kubernetes 1.31.
For a list of available versions, see the Versions subsection of the Managed Kubernetes Product Description instructions ;
-
added the ability to create Managed Kubernetes clusters on dedicated GPU servers;
-
added the ability to specify a CIDR for a node group subnet in a Managed Kubernetes cluster on dedicated servers when creating a cluster and adding a node group;
-
added information about the network plug-in (CNI) that you selected when creating the cluster in the dashboard on the cluster page.
- interruptible node groups are now available in the regions of St. Petersburg, Moscow and Novosibirsk;
- added a field to the cluster deletion form to specify the reason for deleting a cluster.