Skip to main content
Version: Weekly Build

Virtual Kubernetes Cluster

The Virtual Kubernetes Cluster provider gives a resource group a managed Kubernetes control plane inside Codesphere for cloud-native workloads, Helm-based deployments, and advanced orchestration scenarios. In the current provider definitions, this provider is a preview Advanced Compute service with a single customizable plan.

PropertyValueNotes
Provider namevirtual-k8sUse this in landscape provider definitions.
Versionv1Current schema version exposed by the provider.
CategoryAdvanced ComputeShown in the managed services catalog.
ScopeglobalAvailable at resource group scope rather than being tied to a single workspace runtime.
Resource group singletontrueA resource group can only have one Virtual Kubernetes Cluster service instance.
Pause supportfalseThis provider does not support pausing.
Preview Feature

This provider is currently a preview feature. It is not enabled by default and needs to be enabled by your operator.

Schema​

This provider does not expose provider-specific config. Access is handled through the injected kubeconfig and Codesphere networking.

Plan​

The provider exposes one plan, Custom (id: 0), which allows all resource limits to be adjusted within the provider-defined ranges. Example plan: Custom (id: 0).

Limits are configurable

The following table shows the default limits. For Private Cloud Installations, they can be adjusted on cluster level. Consult your Operator for the allowed ranges.

ParameterTypeDefaultMinimumMaximumStaticNotes
cpuinteger2020160NovCPU limit, priced as cpu-tenths.
memoryinteger5120512032768NoRAM limit in MiB, priced as ram-mib.
storageinteger2000020000120000NoPersistent storage limit in MiB, priced as storage-mib.
ephemeralStorageinteger3000030000120000NoEphemeral storage limit in MiB, priced as storage-mib.

Each parameter is billed according to its configured value, not actual consumption — see How billing works for a worked example. ephemeralStorage defaults to a generous 30 GiB (30000 MiB); consider lowering it if your workload doesn't need that much scratch space.

Other runtimes typically access this provider indirectly. Workspaces and landscapes use the injected ~/.kube/config to deploy into the cluster with kubectl or helm, while services running inside the cluster can access other Codesphere resources over the same resource group network. For the runtime-specific URL patterns and kubeconfig workflow, see Virtual Clusters.