Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.cnap.tech/llms.txt

Use this file to discover all available pages before exploring further.

A cluster is a group of one or more machines that are managed by Kubernetes to deploy your applications and keep them running. Clusters are where you run applications for your team, deploy products to customers, and host your software. CNAP provides where CNAP manages the and handles all Kubernetes operations on your (joining, configuring, maintaining), while you stay in control of your infrastructure and billing.
clusters overview

​
Choosing Your Cluster Type

CNAP offers two ways to set up your clusters: managed clusters where CNAP handles all Kubernetes operations, or imported clusters where you bring your existing Kubernetes infrastructure. Both approaches let you deploy applications for your team, sell products to customers, and host your software, but managed clusters remove the operational complexity of running Kubernetes yourself.

​
Managed Clusters

Managed clusters are fully provisioned and maintained by CNAP. CNAP’s provides fully managed infrastructure where CNAP manages the control plane and handles all Kubernetes operations on your workers, while you maintain control over your infrastructure choices and billing. You add your own workers to the cluster, and CNAP handles joining, configuring, and maintaining them. This option is ideal for most users who want to focus on building and selling software without the operational overhead of managing Kubernetes.
Managing a Kubernetes control plane requires expertise in multiple complex areas. You need to handle certificate management and rotation to keep your cluster secure, manage etcd cluster operations including backup and recovery to protect your data, configure the API server with proper security settings, tune the controller manager and scheduler for optimal performance, set up network policies and CNI configuration for proper networking, and establish monitoring, logging, and observability to track cluster health.The control plane also needs dedicated infrastructure including an extra IP address, load balancing, and failover mechanisms to ensure high availability. All of this is infrastructure complexity that stands between you and running your applications.CNAP handles all of this for you, so you can focus on building and selling software instead of managing infrastructure operations.
Key benefits:
  • Zero operational overhead - CNAP manages the control plane
  • Automated provisioning - Get from zero to production-ready cluster in minutes
  • Self-healing infrastructure - Automatic recovery from component failures
  • 24/7 monitoring - Proactive issue detection and resolution
  • Automatic updates - Security patches and Kubernetes version upgrades handled for you
  • Security hardening - Built-in security best practices and regular updates
  • High availability - Multi-zone control plane for maximum uptime
  • Multi-cloud support - Use any provider or mix providers
  • Streamlined onboarding - Developer-friendly experience reduces time-to-market
  • Cost optimization - Pay only for compute resources you actually use
  • Seamless scaling - Add or remove workers as your needs change
Learn how to create a managed cluster →

​
Imported Clusters

Imported clusters allow you to connect your existing Kubernetes infrastructure from any provider (EKS, GKE, AKS, self-managed, or others). This is suitable if you already have an investment in Kubernetes, require specific custom configurations, or prefer to maintain your current cluster management approach.
  • Any Kubernetes cluster - EKS, GKE, AKS, self-managed, or others
  • Cluster management - Control plane and worker operations are managed by your cloud provider or yourself, depending on your cluster type
  • CNAP handles product deployments - CNAP deploys products to your imported cluster, but cluster operations remain outside CNAP’s management
  • Existing investment - Leverage current Kubernetes infrastructure
Learn how to import a cluster →
New to Kubernetes? CNAP’s KaaS is designed to be beginner-friendly while remaining powerful enough for enterprise workloads. CNAP’s managed approach removes the learning curve typically associated with Kubernetes operations.
KaaS seamlessly integrates with all other CNAP features including marketplace deployment, user management, billing, and monitoring. It’s designed to provide a complete platform experience rather than just infrastructure.

​
How Clusters Work with Products

CNAP automatically handles product deployments when customers complete purchases. The platform synchronizes your product’s Helm chart resources with the destination cluster based on the customer’s selected region. Each product deployment gets its own dedicated to avoid conflicting resources and provide security and isolation between different customer deployments. When deployments fail or customers cancel, CNAP automatically cleans up all associated resources to keep your infrastructure tidy.
Resource strategy: Use shared clusters when you control the software. For customers running custom code (container images, game server plugins), use dedicated machines or clusters. Alternatively, use container runtimes like gVisor or Firecracker VMs for stronger isolation without creating new clusters.

​
Cluster suspension

Managed clusters with no attached compute are automatically suspended to save resources. When suspended, the cluster’s control plane is paused — it consumes no compute resources but all configuration, certificates, and state are preserved on disk.

​
What happens during suspension

  • The control plane stops running (no API server, no scheduler)
  • All cluster data is preserved — nothing is deleted
  • The cluster appears as Suspended in your dashboard
  • You are not charged for control plane compute while suspended

​
Suspending a cluster

You can manually suspend a running cluster from the dashboard or via the API: 
curl -X POST -H "Authorization: Bearer $CNAP_TOKEN" \
  -H "X-Workspace-Id: $WORKSPACE_ID" \
  https://api.cnap.tech/v1/clusters/YOUR_CLUSTER_ID/suspend
A cluster can only be suspended when it has no active machines. If your cluster has running machines, delete or suspend them first.

​
Resuming a cluster

When you request a machine for a suspended cluster (through the dashboard or API), CNAP automatically resumes the cluster before provisioning. This takes approximately 15-25 seconds. You can also manually resume a cluster: 
curl -X POST -H "Authorization: Bearer $CNAP_TOKEN" \
  -H "X-Workspace-Id: $WORKSPACE_ID" \
  https://api.cnap.tech/v1/clusters/YOUR_CLUSTER_ID/resume

​
Auto-suspend

Free-tier clusters are automatically suspended when all of these conditions are true:
  • No CNAP-managed machines attached to the cluster
  • No healthy worker nodes in the cluster (nodes disconnected for more than 15 minutes are considered inactive)
CNAP checks these conditions every 15 minutes. Pro-tier clusters are not automatically suspended.
Suspension preserves everything about your cluster: installed applications, RBAC rules, secrets, and configuration. When you resume the cluster, it picks up exactly where it left off.