Manifests & kubectl

Examples

With CloudBroker and Tailscale running, apply the CRDs, then these manifests. Each example includes NodePool, NodeClass, workload YAML, and exact kubectl commands. Replace placeholders with your values.

Prerequisites
  • CloudBroker running at http://localhost:8000
  • Tailscale installed
  • Provider credentials in Kubernetes secrets
  • Replace placeholders in each example: <HOST_TAILSCALE_IP>, project ID, etc.

Provider examples

Google Cloud Platform NodePool, NodeClass, and workload with nodeAffinity for GCP. Full YAML and kubectl commands. View example → Amazon Web Services NodePool, NodeClass (AMI, subnet, credentials), and workload targeting AWS nodes. View example → Microsoft Azure NodePool, NodeClass (subscription, resource group, subnet), and workload for Azure. View example → Hetzner Cloud NodePool, NodeClass (location, API token), and workload for Hetzner. View example → Scaleway NodePool, NodeClass (zone, project, API key), and workload for Scaleway. View example → DigitalOcean NodePool, NodeClass (region, API token), and workload for DigitalOcean. View example → OVH Cloud NodePool, NodeClass (project, region, API credentials), and workload for OVH. View example → Aruba Cloud NodePool, NodeClass (project ID, API credentials), and workload for Aruba Cloud. View example →

Advanced examples

Kind Kind cluster integrations Index: Tailscale Operator, Local Path Provisioner, Longhorn. Links to each integration. View index → Tailscale Tailscale Operator Expose services to tailnet via LoadBalancer or Ingress. No public IPs or load balancers. View example → Storage Local Path Provisioner Simple local storage on each node. Dev/test and single-node workloads. View example → Storage Longhorn Distributed replicated storage. HA, snapshots, GUI. Combined example with Tailscale + pod affinity. View example → CNPG CloudNativePG Distributed PostgreSQL with CloudNativePG. Restore replicas on newly provisioned burst nodes via streaming replication. View example → Flex Workload without nodeAffinity Pod without nodeSelector/nodeAffinity can land on any burst node. Full manifest and commands. View example → DR Disaster recovery Multi-provider NodePool, automatic failover, simulation script. NodePool + NodeClass + Deployment. View example → Scale Multiple unscheduled pods Demand aggregation: 5 pods trigger one larger node. Deployment YAML and kubectl commands. View example → Run Apply & results Step-by-step: install CRDs, create secrets, apply manifest. kubectl get output examples. View example → Labels Pod labels and affinities Index: nodeAffinity, podAffinity (co-locate), podAntiAffinity (spread). Links to each example. View index → Node Node affinity No affinity, provider affinity (GCP), nodepool affinity. Target specific clouds or NodePools. View example → Pod Pod affinity (co-locate) Co-locate related pods on the same node to avoid egress fees. View example → Pod Pod anti-affinity (spread) Spread replicas across nodes for high availability. View example →

Explore further

Articles → Architecture → Quickstart →