Manifests & kubectl

Examples

With CloudBroker and Tailscale running, apply the CRDs, then these manifests. A high-resource pod triggers provisioning; kubectl get shows the result. Replace placeholders with your values.

GCP #

NodePool + NodeClass + Workload (nodeAffinity: gcp)

# NodePool
apiVersion: cloudburst.io/v1alpha1
kind: NodePool
metadata:
  name: gcp-nodepool
  namespace: default
spec:
  requirements:
    regionConstraint: "ANY"
    arch: ["x86_64"]
    maxPriceEurPerHour: 0.15
    allowedProviders: ["gcp"]
  limits:
    maxNodes: 3
    minNodes: 0
  template:
    labels:
      cloudburst.io/nodepool: "gcp-nodepool"
      cloudburst.io/provider: "gcp"
  disruption:
    ttlSecondsAfterEmpty: 60
  weight: 1
---
# NodeClass
apiVersion: cloudburst.io/v1alpha1
kind: NodeClass
metadata:
  name: gcp-nodeclass
  namespace: default
spec:
  gcp:
    project: "your-gcp-project-id"
    zone: "us-central1-a"
  join:
    hostApiServer: "https://<HOST_TAILSCALE_IP>:6443"
    kindClusterName: "cloudburst"
    tokenTtlMinutes: 60
  tailscale:
    authKeySecretRef:
      name: tailscale-auth
      key: authkey
  bootstrap:
    kubernetesVersion: "1.34.3"
---
# Workload (triggers burst; targets GCP nodes)
apiVersion: v1
kind: Pod
metadata:
  name: gcp-workload
  namespace: default
spec:
  containers:
  - name: workload
    image: busybox:1.36
    command: ["sleep", "infinity"]
    resources:
      requests:
        cpu: "1500m"
        memory: "2Gi"
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: cloudburst.io/provider
            operator: In
            values: ["gcp"]

↑ Back to menu

AWS #

NodePool + NodeClass + Workload (nodeAffinity: aws)

# NodePool
spec:
  requirements:
    allowedProviders: ["aws"]
  template:
    labels:
      cloudburst.io/nodepool: "aws-nodepool"
      cloudburst.io/provider: "aws"
---
# NodeClass (provider section)
spec:
  aws:
    region: "eu-west-1"
    ami: "ami-xxxxxxxx"
    subnetID: "subnet-xxxxxxxx"
  join: { hostApiServer: "...", kindClusterName: "cloudburst", tokenTtlMinutes: 60 }
  tailscale: { authKeySecretRef: { name: tailscale-auth, key: authkey } }
  bootstrap: { kubernetesVersion: "1.34.3" }
---
# Workload
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: cloudburst.io/provider
            operator: In
            values: ["aws"]

↑ Back to menu

Azure #

NodePool + NodeClass + Workload (nodeAffinity: azure)

# NodePool
spec:
  requirements:
    allowedProviders: ["azure"]
  template:
    labels:
      cloudburst.io/nodepool: "azure-nodepool"
      cloudburst.io/provider: "azure"
---
# NodeClass (provider section)
spec:
  azure:
    subscriptionID: "your-subscription-id"
    resourceGroup: "your-resource-group"
    location: "westeurope"
    subnetID: "/subscriptions/.../subnets/your-subnet"
    clientIDSecretRef: { name: azure-credentials, key: AZURE_CLIENT_ID }
    clientSecretSecretRef: { name: azure-credentials, key: AZURE_CLIENT_SECRET }
  join: { hostApiServer: "...", kindClusterName: "cloudburst", tokenTtlMinutes: 60 }
  tailscale: { authKeySecretRef: { name: tailscale-auth, key: authkey } }
  bootstrap: { kubernetesVersion: "1.34.3" }
---
# Workload
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: cloudburst.io/provider
            operator: In
            values: ["azure"]

↑ Back to menu

Hetzner #

NodePool + NodeClass + Workload (nodeAffinity: hetzner)

# NodePool
spec:
  requirements:
    allowedProviders: ["hetzner"]
  template:
    labels:
      cloudburst.io/nodepool: "hetzner-nodepool"
      cloudburst.io/provider: "hetzner"
---
# NodeClass (provider section)
spec:
  hetzner:
    location: "fsn1"
    apiTokenSecretRef: { name: hetzner-credentials, key: HETZNER_API_TOKEN }
    image: "ubuntu-22.04"
  join: { hostApiServer: "...", kindClusterName: "cloudburst", tokenTtlMinutes: 60 }
  tailscale: { authKeySecretRef: { name: tailscale-auth, key: authkey } }
  bootstrap: { kubernetesVersion: "1.34.3" }
---
# Workload
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: cloudburst.io/provider
            operator: In
            values: ["hetzner"]

↑ Back to menu

Scaleway #

NodePool (Scaling policy: EU, x86_64, max €0.15/h, Scaleway)

apiVersion: cloudburst.io/v1alpha1
kind: NodePool
metadata:
  name: scaleway-nodepool
  namespace: default
spec:
  requirements:
    regionConstraint: "EU"
    arch: ["x86_64"]
    maxPriceEurPerHour: 0.15
    allowedProviders: ["scaleway"]
  limits:
    maxNodes: 3
    minNodes: 0
  template:
    labels:
      cloudburst.io/nodepool: "scaleway-nodepool"
      cloudburst.io/provider: "scaleway"
  disruption:
    ttlSecondsAfterEmpty: 60
  weight: 1

NodeClass (Scaleway zone, join config, Tailscale auth ref)

apiVersion: cloudburst.io/v1alpha1
kind: NodeClass
metadata:
  name: scaleway-nodeclass
  namespace: default
spec:
  scaleway:
    zone: "fr-par-1"
    projectID: "your-scaleway-project-id"
    image: "ubuntu_jammy"
    apiKeySecretRef:
      name: scaleway-credentials
      key: SCW_SECRET_KEY
  join:
    hostApiServer: "https://<HOST_TAILSCALE_IP>:6443"
    kindClusterName: "cloudburst"
    tokenTtlMinutes: 60
  tailscale:
    authKeySecretRef:
      name: tailscale-auth
      key: authkey
  bootstrap:
    kubernetesVersion: "1.34.3"

Workload (Pod that can't be scheduled — triggers burst)

apiVersion: v1
kind: Pod
metadata:
  name: scaleway-workload
  namespace: default
spec:
  containers:
  - name: workload
    image: busybox:1.36
    command: ["sleep", "infinity"]
    resources:
      requests:
        cpu: "1500m"
        memory: "2Gi"
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: cloudburst.io/provider
            operator: In
            values: ["scaleway"]

↑ Back to menu

DigitalOcean #

NodePool + NodeClass + Workload (nodeAffinity: digitalocean)

# NodePool
spec:
  requirements:
    allowedProviders: ["digitalocean"]
  template:
    labels:
      cloudburst.io/nodepool: "digitalocean-nodepool"
      cloudburst.io/provider: "digitalocean"
---
# NodeClass (provider section)
spec:
  digitalocean:
    region: "ams3"
    apiTokenSecretRef: { name: digitalocean-credentials, key: DIGITALOCEAN_API_TOKEN }
    image: "ubuntu-22-04-x64"
  join: { hostApiServer: "...", kindClusterName: "cloudburst", tokenTtlMinutes: 60 }
  tailscale: { authKeySecretRef: { name: tailscale-auth, key: authkey } }
  bootstrap: { kubernetesVersion: "1.34.3" }
---
# Workload
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: cloudburst.io/provider
            operator: In
            values: ["digitalocean"]

↑ Back to menu

OVH #

NodePool + NodeClass + Workload (nodeAffinity: ovh)

# NodePool
spec:
  requirements:
    allowedProviders: ["ovh"]
  template:
    labels:
      cloudburst.io/nodepool: "ovh-nodepool"
      cloudburst.io/provider: "ovh"
---
# NodeClass (provider section)
spec:
  ovh:
    projectID: "your-ovh-project-id"
    region: "GRA"
    applicationKeySecretRef: { name: ovh-credentials, key: OVH_APPLICATION_KEY }
    applicationSecretSecretRef: { name: ovh-credentials, key: OVH_APPLICATION_SECRET }
    consumerKeySecretRef: { name: ovh-credentials, key: OVH_CONSUMER_KEY }
  join: { hostApiServer: "...", kindClusterName: "cloudburst", tokenTtlMinutes: 60 }
  tailscale: { authKeySecretRef: { name: tailscale-auth, key: authkey } }
  bootstrap: { kubernetesVersion: "1.34.3" }
---
# Workload
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: cloudburst.io/provider
            operator: In
            values: ["ovh"]

↑ Back to menu

Workload without nodeAffinity #

Pods without nodeSelector or nodeAffinity can land on any burst node with capacity. When multiple NodePools exist, the scheduler may place the pod on any provider. Use this when you don't care which cloud hosts the workload.

apiVersion: v1
kind: Pod
metadata:
  name: any-provider-workload
  namespace: default
spec:
  containers:
  - name: workload
    image: busybox:1.36
    command: ["sleep", "infinity"]
    resources:
      requests:
        cpu: "1500m"
        memory: "2Gi"

↑ Back to menu

Apply & results #

Apply manifests

kubectl apply -f config/crd/bases/
kubectl apply -f config/samples/nodeclass-scaleway.yaml
kubectl apply -f config/samples/nodepool-scaleway.yaml
kubectl apply -f config/samples/workload-scaleway.yaml

Result of kubectl apply

customresourcedefinition.apiextensions.k8s.io/nodeclaims.cloudburst.io created
customresourcedefinition.apiextensions.k8s.io/nodeclasses.cloudburst.io created
customresourcedefinition.apiextensions.k8s.io/nodepools.cloudburst.io created
nodeclass.cloudburst.io/scaleway-nodeclass created
nodepool.cloudburst.io/scaleway-nodepool created
pod/scaleway-workload created

Result of kubectl get nodeclaims

NAME                    NODEPOOL            PROVIDER   REGION     INSTANCE TYPE   PHASE     AGE
scaleway-nodepool-xxxx   scaleway-nodepool   scaleway   fr-par-1   DEV1-M          Running   2m15s

Result of kubectl get nodes -l cloudburst.io/provider=scaleway

NAME                          STATUS   ROLES    AGE   VERSION
scaleway-nodepool-xxxx-xxxxx   Ready    <none>   2m    v1.34.3

↑ Back to menu

Pod labels and NodePool matching #

Pods do not need to specify cloudburst.io/nodepool or cloudburst.io/provider. These labels are applied to nodes from the NodePool template, not required on pods.

Demand detection

Cloudburst treats any Pending pod with reason Unschedulable as demand (except kube-system and pods annotated cloudburst.io/ignore=true). Each NodePool sees all unschedulable pods and may create capacity. There is no per-pod filtering by NodePool or provider labels.

Scheduling behavior

When a NodePool provisions a node, that node gets the NodePool's template labels (e.g. cloudburst.io/nodepool, cloudburst.io/provider). For the scheduler to place a pod on it:

  • No nodeSelector/nodeAffinity — the pod can land on any node, including burst nodes.
  • Matching nodeSelector/nodeAffinity — the pod will only land on nodes that satisfy it (e.g. cloudburst.io/provider: gcp).
  • Non-matching affinity — if the pod requires cloudburst.io/provider: gcp but the node is from a Scaleway NodePool, the pod stays unschedulable until a matching node exists.

Multiple NodePools

With several NodePools (e.g. GCP, Scaleway, Hetzner), each sees the same global unschedulable demand and may create nodes independently. Which node the pod lands on depends on which nodes match the pod's affinity (if any) and scheduler decisions. Use nodeSelector or nodeAffinity on pods when you want to target specific providers or NodePools.

No nodeSelector/nodeAffinity

Pod can be scheduled on any burst node with capacity.

cloudburst.io/provider: gcp

Pod will only land on nodes from a NodePool whose template includes that label (typically GCP).

cloudburst.io/nodepool: my-pool

Pod will only land on nodes created by that NodePool.

↑ Back to menu

Secrets (Tailscale auth key, provider API keys) go in Kubernetes secrets; reference them from the NodeClass. Full manifests in the repo config/samples/.

Use cases

When Cloudburst fits best. Full use cases on the homepage.

Cloudburst vs alternatives

Full comparison table on the homepage.

Got questions?

Full FAQs on the homepage.