Cluster Components

Reference for the components within the CKS Managed Control Plane and Managed Data Plane

Kubernetes versions

CoreWeave Kubernetes Service (CKS) supports multiple Kubernetes versions to ensure compatibility with a wide range of applications and workloads. This allows customers to choose a version that best fits their needs while also ensuring access to the latest features and security updates.

Supported versions

CKS can create clusters with these Kubernetes versions, which use the control group (cgroup) version listed.

Kubernetes Version	Control Group Version	Notes
v1.30	cgroup v1
v1.31	cgroup v1
v1.32	cgroup v1	Default version for new clusters
v1.33	cgroup v2

Deprecated versions

These versions are no longer supported for new CKS clusters. Existing clusters can continue to operate until they are upgraded or decommissioned.

Kubernetes Version	Control Group Version	Notes
v1.29	cgroup v1	Deprecated

Managed Control Plane

To ensure security and stability, CoreWeave provides a Managed Control Plane for CKS clusters. Learn more about CKS cluster architecture.

Managed Data Plane components

The CKS Managed Data Plane comes with several preinstalled components, which allow customers to start running workloads immediately after deploying their CKS cluster. These components are configurable to allow for the desired state of a CKS cluster.

Info

• Some older clusters' namespaces may not have the cw- prefix shown in the tables below.
• Components and API versions change often. The versions deployed to your cluster may be newer than shown below.

Networking

Component	Namespace	Version	Description
cilium	cw-cilium-system	1.16.1	Container Network Interface (CNI) that uses eBPF for efficient load-balancing.
calico (deprecated)	cw-calico-system	3.29.1	Container Network Interface (CNI) used on older clusters; no longer supported for new CKS clusters.
multus-cni			Enables attaching multiple network interfaces to Pods. Used to configure InfiniBand as a secondary interface for workloads that require high throughput and low latency in HPC environments.
coredns		1.11.1	A flexible DNS server consisting of multiple plugins that manage in-cluster DNS and resolve service records inside clusters. This runs in the CKS Managed Control Plane.
konnectivity-agent	kube-system	0.30.3	Isolates CKS Control Plane traffic from the cluster network and creates a TCP level proxy to konnectivity-server, which sits alongside the Control Plane to tunnel traffic from the Control Plane to in-cluster endpoints.
metallb	cw-metallb-system	0.13.4	Handles IP Address Management (IPAM) for LoadBalancer services; is not used for advertisements. See How to: Expose a Service
node-local-dns	kube-system	1.22.28	Improves cluster DNS performance and speeds up traffic between CoreDNS instances. The node-local-dns Pods run CoreDNS in local caching mode as a DaemonSet over TCP, avoiding dropped packets and reducing tail latency.

Networking notes

• Cilium, like CKS, is a managed product. We deploy Cilium with a minimal feature set. If you require a feature that isn't enabled, or is only available in Cilium Enterprise, please contact support.
• Argo CD has documented issues with Cilium when not using annotation tracking. To prevent issues, declare resource exclusions in the Argo CD ConfigMap.
• Customers who require a specific CNI (such as Cilium Enterprise instead of Cilium Open Source) can bring their own. This manual process requires coordination with CoreWeave when deploying the cluster. Please contact support for more information.

Cluster management

Component	Namespace	Version	Description
hpc-verification	cw-hpc-verification	1.8.0	Validates the Node is in an optimal state. These tests only run on idle Nodes.
node-problem-detector	cw-node-problem-detector	1.5.0	Framework for running health checks for each Node run as a DaemonSet. It allows CoreWeave to quickly identify Nodes that are misbehaving and mark them as Unschedulable.
nfd	cw-nfd	0.9.0	Node Feature Discovery. Detects hardware features and labels Nodes accordingly for use in scheduling decisions.

Info

Older clusters may have jspolicy installed in the jspolicy-policies-<Org ID>-<cluster ID> and tenant-<Org ID>-<cluster ID> namespaces. Newer clusters do not include JSPolicy.

Observability and security

Component	Namespace	Version	Description
victoria-metrics	cw-victoria-metrics	0.45.0	Collects metrics from inside the cluster and relays them to a logging solution.
vmagent	cw-victoria-metrics		Gathers metrics for Victoria Metrics.
metrics-server	kube-system	0.6.2	Only installed on clusters deployed before July 7, 2025.
node-pci-exporter	cw-exporters	1.5.0	Extends the node-exporter running as part of the Prometheus stack, exposing additional metrics about deployed GPUs.
gpu-thermal-exporter	cw-exporters		Emits hardware metrics for GPU thermal data for Node health monitoring.
nvlink-exporter	cw-exporters		Exports metrics for NVLink performance in GPU Nodes.
ping-exporter	cw-ping-exporter		Measures packet loss to configured network targets for connectivity checks.
promtail	cw-promtail		Forwards Node logs centrally using Grafana's Promtail.
crowdstrike	cw-crowdstrike	7.19.0-17219-1.falcon-linux.Release.US-2	CrowdStrike Falcon cybersecurity tooling.

Storage

Component	Namespace	Version	Description
vast-csi	cw-vast-csi	2.3.1	Container Storage Interface for Vast Storage.

Object Storage

Component	Namespace	Version	Description
lota	cw-object-storage	v1.25.1	CoreWeave's Local Object Transfer Accelerator (LOTA) for AI Object Storage. Learn more about LOTA.

Hardware

Component	Namespace	Version	Description
k8s-device-plugin	cw-nvidia-device-plugin	0.12.3	NVIDIA's official device plugin for GPU-enabled containers in Kubernetes.
nvidia-imex	cw-nvidia-imex		Enables the NVIDIA Internode Memory Exchange (IMEX) service.
rdma-shared-dp	cw-rdma		Device plugin to support InfiniBand and RDMA hardware resources.

Warning

CKS does not require NVIDIA's GPU Operator. To avoid conflicts, customers should not attempt to install it.