name | required | Unique cluster name |
description | | Free text shown with the cluster |
environment | development | What the cluster runs: production, staging, development, testing, local, or a name of your own. The portal asks for it as What will this cluster run? |
criticality | the environment’s own level | How important the cluster is: production, staging, development or disposable. Also sets the default load_balancer_plan |
credential_id | required | UpCloud API credential ID |
ssh_key_credential_id | required | SSH key credential ID |
zone | required | UpCloud datacenter zone |
zones | | Zone pool for a multi-zone cluster, must include zone (the primary). At least 3 zones and 3 control planes; kubeadm only. See Multi-zone clusters |
network_mode | derived | private_network or wireguard_mesh; wireguard_mesh for a pool, and on a single zone it makes the cluster mesh-capable so zones can be added later. On a wireguard_mesh cluster the load balancer backends are addressed on the private network rather than the overlay, so LoadBalancer Services work unchanged - see the guide |
network_ip_range | a /24 picked for the cluster | Private network IP range. The portal accepts a private range of /24 or larger. Left out, Ankra picks a 10.x.y.0/24 for this cluster, different for each cluster. A multi-zone pool takes one consecutive /24 per zone from this range |
bastion_plan | STARTER-1xCPU-1GB | Server plan for the bastion host |
bastion_allowed_ips | any address | IPv4 addresses or CIDRs allowed to reach the bastion over SSH, at most 64. IPv6 and 0.0.0.0/0 are refused; leave it out to keep SSH open to any address (key-only). Ankra’s own egress is always allowed. Change it later with ankra cluster upcloud bastion allowed-ips |
gateway_plan | standard | Plan of the NAT gateway all node egress goes through. List the plans with ankra cluster upcloud gateway-plans; change it later with ankra cluster upcloud gateway-plan, which cuts egress for several minutes. See NAT gateway and load balancer plans |
load_balancer_plan | production-small for a production cluster, development otherwise | Plan of the UpCloud load balancer in front of the ingress Ankra installs. Needs include_networking. List the plans with ankra cluster upcloud load-balancer-plans; change it later with ankra cluster upcloud load-balancer-plan set |
control_plane_count | 1 | Number of control plane nodes (1 or 3) |
control_plane_plan | PREMIUM-4xCPU-8GB | Server plan for control planes |
worker_count | 1 | Number of worker nodes (legacy, use node_groups instead) |
worker_plan | PREMIUM-2xCPU-4GB | Server plan for workers (legacy, use node_groups instead) |
node_groups | | Array of node group definitions (see Node groups) |
distribution | kubeadm | Kubernetes distribution (kubeadm or k3s) |
kubernetes_version | latest stable | Kubernetes version (optional) |
k3s_disabled_components | | k3s only. Packaged components the k3s install leaves out (k3s --disable), on top of what the provider’s k3s shape already disables: any of coredns, servicelb, traefik, local-storage, metrics-server, runtimes, for example ["traefik"] on a cluster that brings its own ingress. A list on a kubeadm cluster is refused with a 400 (k3s_disabled_components is only supported for the k3s distribution.); an entry outside those six or an entry listed twice is a 422 validation error pointing at the entry |
cni | cilium (kubeadm), flannel (k3s) | CNI plugin: flannel, calico or cilium. Follows the distribution when omitted; kubeadm clusters always use cilium |
cni_features | all off | CNI feature toggles, fixed at creation. See Choices fixed at create time |
etcd_topology | stacked | kubeadm only. stacked or external (dedicated etcd servers) |
etcd_node_count | 3 | kubeadm external topology only (3 or 5) |
etcd_plan | PREMIUM-4xCPU-8GB | kubeadm external topology only. Server plan for the etcd nodes |
external_cloud_provider | true | Install the UpCloud cloud controller manager and CSI driver and wire the nodes for them |
include_networking | true | Deploy the networking stack: Traefik, cert-manager and a letsencrypt-prod Let’s Encrypt ClusterIssuer |
gateway_api | false | Also serve the Kubernetes Gateway API from the networking stack’s Traefik. Needs include_networking. See Gateway API |
include_dns | true | Give the cluster a delegated subdomain on ankra.cc with external-dns wired |
gitops_provider | github | Git provider of the GitOps repository: github or bitbucket_cloud |
gitops_credential_name | | Name of the Git credential for GitOps integration |
gitops_repository | | GitOps repository, for example my-org/my-repo |
gitops_workspace | | Bitbucket Cloud only. The workspace that holds the repository |
gitops_repo_slug | | Bitbucket Cloud only. The repository slug |
gitops_branch | master | Branch for GitOps pushes |