Cluster API is the Kubernetes sub-project for declarative cluster lifecycle management. Used by platform teams to provision and manage workload clusters.Documentation Index
Fetch the complete documentation index at: https://radarhq.io/docs/llms.txt
Use this file to discover all available pages before exploring further.
What Radar Shows
Topology: Full CAPI ownership chain - ClusterClass → Cluster → KubeadmControlPlane → Machine → Node, and Cluster → MachineDeployment → MachineSet → Machine → Node. MachineHealthCheck → Cluster protection edges. Machine → Node edges use status.nodeRef (semantic, not owner-ref). Cluster Detail View:- Phase, version, cluster class, control plane endpoint
- Control plane and worker replica counts (v1beta2-aware)
- Control plane and infrastructure references (clickable)
- ClusterClass topology section (worker MachineDeployments table)
- “Connect to Cluster” button - auto-connects Radar to the workload cluster
- “Download Kubeconfig” button
- Conditions
- Phase, role (Control Plane / Worker), version, provider ID
- Clickable Node reference (via status.nodeRef)
- Addresses table, node info (OS, architecture, kernel, kubelet)
- Bootstrap and infrastructure references
- Phase, replicas (desired/ready/available/up-to-date), strategy
- Version, cluster name
- Machine template references
- Owned machines label hint (copyable)
- Replicas, version, initialized status (v1beta2-aware)
- Machine template with drain/volume detach/deletion timeouts
- Kubeadm config highlights (cert SANs)
- Last remediation info
- Owned machines label hint
- Infrastructure, control plane, worker topology tables
- Variables with schema types
- Patches with definitions and enabledIf expressions
- Expected/healthy machine counts, remediations allowed
- Label selector display
- Unhealthy conditions tables (v1beta1 + v1beta2 formats)
- Remediation template
topology.cluster.x-k8s.io/owned) show a warning banner.
Fleet topology mode: Dedicated “Fleet” view filters to CAPI and infrastructure provider resources only, giving a clean cluster-management view without application workload noise. Groups start expanded by default.
Infrastructure Provider Renderers
Radar has first-class renderers for AWS (CAPA), GCP (CAPG), and Azure (CAPZ) infrastructure provider resources. These surface provider-specific operational data - instance types, scaling config, VPC/subnet topology, managed service addons - that would otherwise be buried in raw YAML. AWS EKS control plane - VPC topology with subnets (Public/Private badges), security groups, EKS addons, IAM roles:
Supported CRDs
| CRD | Group | Topology | Detail View | AI Summary |
|---|---|---|---|---|
| Cluster | cluster.x-k8s.io | Yes | Yes | Yes |
| ClusterClass | cluster.x-k8s.io | Yes | Yes | Yes |
| Machine | cluster.x-k8s.io | Yes | Yes | Yes |
| MachineSet | cluster.x-k8s.io | Yes | Yes | Yes |
| MachineDeployment | cluster.x-k8s.io | Yes | Yes | Yes |
| MachinePool | cluster.x-k8s.io | Yes | Yes | Yes |
| MachineHealthCheck | cluster.x-k8s.io | Yes | Yes | Yes |
| MachineDrainRule | cluster.x-k8s.io | No | Yes | No |
| KubeadmControlPlane | controlplane.cluster.x-k8s.io | Yes | Yes | Yes |
| KubeadmControlPlaneTemplate | controlplane.cluster.x-k8s.io | No | Generic | No |
| KubeadmConfig | bootstrap.cluster.x-k8s.io | No | Yes | No |
| KubeadmConfigTemplate | bootstrap.cluster.x-k8s.io | No | Generic | No |
| AWSManagedControlPlane | controlplane.cluster.x-k8s.io | Yes | Yes | No |
| AWSManagedMachinePool | infrastructure.cluster.x-k8s.io | Yes | Yes | No |
| AWSMachine | infrastructure.cluster.x-k8s.io | Yes | Yes | No |
| AWSMachineTemplate | infrastructure.cluster.x-k8s.io | No | Yes | No |
| AWSManagedCluster | infrastructure.cluster.x-k8s.io | No | Yes | No |