•  
    •  
  • website icon
My Account
Support
Kubernetes clusters
Deploy a production‑ready Kubernetes cluster in your DevSpace: with a fault‑tolerant control plane, node autoscaling, and public IPs for your services. You’re working with standard Kubernetes — the standard API, the standard kubectl, no proprietary wrappers.
A full‑fledged managed Kubernetes — in minutes, not weeks.
Free Trial
What you get
  • -1-
    Cluster on request

    Creating it from the control panel in a few steps, without requests or waiting.

  • -2-
    Vanilla Kubernetes

    Full compatibility with the CNCF ecosystem: Helm charts, operators, and CI/CD tools work without adaptation.

  • -3-
    Auto‑scaling

    cluster-autoscaler adds and removes worker nodes in response to the load.

  • -4-
    Public services out of the box

    A service of the LoadBalancer type automatically receives an external IP address.

  • -5-
    Persistent volumes

    PersistentVolume on the platform’s block storage, including replicated classes.

  • -6-
    Version selection

    Current Kubernetes versions to choose from when creating a cluster.

Free Trial
How the service works
Each cluster is a separate, fully isolated Kubernetes environment within your DevSpace. The CozyStack platform builds it from proven open‑source components:
  • Control plane в контейнерах

    The API server, controller-manager, and scheduler run as managed processes on the platform’s resources — using Kamaji. This is faster and cheaper than dedicated master nodes. The number of control plane replicas is set during creation: several instances run simultaneously, and the failure of one does not interrupt the cluster’s operation.

  • Isolated etcd

    The state of your clusters is stored in a separate etcd instance deployed within your space — it is not shared with other platform clients.

  • The working nodes are virtual machines.

    Workers are created as VMs based on KubeVirt and automatically join the cluster. Node groups with different CPU/memory profiles are designed for different types of workloads.

  • The lifecycle via Cluster API
    The creation, scaling, and updating of nodes are managed declaratively via the Kubernetes Cluster API — the industry standard for cluster management.
  • Network and drives
    External IPs are advertised via BGP and move between nodes during failures. Volumes are provisioned via CSI, and if you select a replicated storage class, a disk or server failure does not mean data loss.
Opportunities
Use cases
  • Microservices in production
    A fault‑tolerant cluster with auto‑scaling and external load balancing; databases and queues for applications are added nearby, from the DevSpace service catalog in the same space, and are accessible via internal addresses — there’s no need to publish them.
  • Dev/test environments
    A separate cluster for each team or feature: created, tested, deleted.
  • CI/CD and GitOps
    The standard Kubernetes API works well with Flux, Argo CD, GitLab CI, and any other familiar tools.
  • Migration from self‑assembled clusters
    Stop manually administering the control plane and etcd; manifests and Helm charts are transferred as is.
Free Trial
How to deploy a Kubernetes cluster in the cloud — three steps
  • 1
    Create a space in the DevSpace panel — an isolated environment for your project.
  • 2
    Add the “Kubernetes” service: select the version, size, and number of worker nodes.
  • 3
    Download the kubeconfig from the cluster page — and work via kubectl, Lens, or CI/CD.

FAQ

Try it out
+7 727 312 3697
info@cloupard.kz


Service Status

© All rights reserved
Client Agreement
WL Partnership Agreement
Appendix 1 “General Rules for Providing Services”
Appendix 2 “Rules for Providing Individual Services” (SLA)
Appendix 3 “Privacy Policy” (NDA)
Appendix 4 “Resale of Services”
Agency agreement
Public offer
Online payment rules