Skip to main content
A cluster’s settings are one page each. Open Infrastructure → Clusters → <your cluster> → Settings (/next/clusters/<cluster-id>/settings): the menu switches to the settings pages, with Terminate or remove at the bottom under Danger zone. Each page below gives its address, what it changes, the permission it needs and the CLI equivalent. Some cluster pages are not settings and sit in the cluster’s own menu: Kubectl access (who may reach the cluster with kubectl, see Cluster access and Kubeconfig), Audit log (who changed what on this cluster, see Audit log) and the nodes (see Manage Nodes and Node Groups).

At a glance

Built-in roles: clusters.write and variables.write are held by Member, Operator, Admin and Owner; clusters.operate by Operator, Admin and Owner; clusters.delete, agents.manage, ai.manage and projects.manage by Admin and Owner. Every page tells you when your role cannot change it, for example “Your role cannot sync this cluster. That needs the clusters.write permission.” See Roles and access.

General

/next/clusters/<cluster-id>/settings Name and use shows the cluster’s Name, Description, Used for and, where it differs, its Label. The name and description are set when the cluster is created or imported, and are not changed here.
  • Change what it’s used for opens a dialog with Used for: Production, Staging, Development or Disposable. Under Change its label you can set a Label, which Cost budgets and Security use to group clusters. Click Review change, then Save. Insight grading, notifications and AI remediation follow what a cluster is used for.
  • Confirm it is <purpose> appears when nobody has chosen the value yet. Confirming records that you chose it; from then on Ankra suggests stopping the cluster at night only if it is not production and is quiet at night.
  • Project shows the project the cluster is in and moves it to another one. It needs projects.manage and appears only when the organisation has projects; see Set up projects.
The portal cannot move a cluster to another organisation. Use ankra cluster move:
You must be an administrator of both organisations. See the CLI reference for what moves with the cluster and what is detached.

Rename the cluster

A cluster can be renamed in place with ankra cluster rename <cluster> <new-name>, without rebuilding it. Server and node names, the DNS zone and the GitOps directory keep the old name. A cluster synced with a connected GitOps repository and an Ankra Cloud cluster cannot be renamed: the platform refuses and changes nothing. Rename a Cluster says what changes and what to update afterwards.

Power schedules

Stops and starts the cluster on a timetable, on the providers that support stop and start. See Stop, Start and Power Schedules.

Agent

/next/clusters/<cluster-id>/settings/agent Shows the agent’s Connection, Version (with Upgrade agent when a newer one is available), Last check-in and Automatic upgrades.
  • Turn automatic upgrades off keeps this cluster’s agent on its version until someone upgrades it by hand; Turn automatic upgrades on lets the platform roll new versions to it. The organisation setting Upgrade cluster agents automatically (/next/organisation/pipelines) must also be on.
  • Upgrade agent upgrades it now. It needs agents.manage.
  • Create an agent install command links to the command on Upgrades and maintenance.
When no agent has registered yet, the page says so: on a cluster you imported, the agent registers once its install command has run. See Cluster agent.

Stop, start and schedules

/next/clusters/<cluster-id>/settings/power Stop or start now stops the cluster’s servers or starts them again, and Schedules does it on a timetable. Only clusters Ankra built or runs at a cloud provider can be stopped. See Stop, Start and Power Schedules.

Upgrades and maintenance

/next/clusters/<cluster-id>/settings/maintenance Automatic patch upgrades of Kubernetes are off by default, and neither the portal nor the CLI has a switch for them.

GitOps

/next/clusters/<cluster-id>/settings/gitops
  • Repository - the Git repository that holds the cluster’s definition. Connect a repository or Change repository.
  • Sync - whether the cluster matches its repository. Ankra syncs when the branch changes; Sync now checks right away.
  • Merge conflicts - resources changed both in Git and in Ankra since the last sync. Ankra does not sync them until you choose which version to keep.
  • Sync history - every sync between the cluster and its repository.
  • Needs you appears when GitOps is not working, for example “Install the Ankra GitHub App”, with what to do.
See GitOps.

Kubernetes sync

/next/clusters/<cluster-id>/settings/kubernetes-sync Ankra keeps a copy of the cluster’s Kubernetes resources so its pages load quickly. Sync status says whether that copy is up to date, and Resource types lists each type Ankra copies. Run a full sync… reads every resource from the cluster again and removes resources from the copy that no longer exist; nothing on the cluster changes. The cluster must be connected.

Network

/next/clusters/<cluster-id>/settings/network
  • On a cluster Ankra builds on UpCloud or OVHcloud, choose the internet gateway and change its size (Change NAT gateway plan, Change load balancer plan or Change gateway size). A gateway change can cut the cluster’s internet access for several minutes, and the dialog says so.
  • On a cluster Ankra built on a cloud provider, Turn the Gateway API on. See Gateway API.
Other clusters have nothing to set here.

SSH access

/next/clusters/<cluster-id>/settings/ssh Connect with SSH lists the cluster’s servers and the commands to Open a shell on the control plane and Reach the Kubernetes API on your computer. SSH keys lists the keys on the servers and changes them. Where SSH is not offered, the page says “SSH access is not offered for this kind of cluster”.

Ankra AI

/next/clusters/<cluster-id>/settings/ai

Logs and data

/next/clusters/<cluster-id>/settings/logs
  • Ankra hosted logs - whether this cluster sends the log lines of its running containers to Ankra, who can read them and who can change the setting. See Hosted logs.
  • Data policy - the data policy managed services on this cluster follow.
To connect a log store you already run (Grafana Loki, Elasticsearch or OpenSearch), use Integrations.

Integrations

/next/clusters/<cluster-id>/settings/integrations
  • Your metrics and logs - Metrics (your own Prometheus, see Prometheus), Log store (Connect a log store, see Log sources) and External log viewer. Each connection can be tested and disconnected.
  • Security and cost reporting - Vulnerability scanning and Cloud cost estimates.
Built-in metrics need none of these; see Cluster metrics. SOPS encryption is set for the whole organisation, not per cluster; see SOPS.

Variables

/next/clusters/<cluster-id>/settings/variables Values the stacks on this cluster can refer to, such as a domain name or a replica count. Reading them needs variables.read; adding, changing and deleting them needs variables.write. See Variables.

Credentials

/next/clusters/<cluster-id>/settings/credentials Credentials this cluster uses lists the credentials the cluster’s definition names: the Git repository it syncs with, its metrics and log sources and the registries its add-ons install from. Open one to check, rotate or replace it. See Credentials.

Terminate or remove

/next/clusters/<cluster-id>/settings/terminate What the page offers depends on the cluster. Every action asks you to type the cluster’s name, and needs clusters.delete: “Your role cannot do this. That needs the clusters.delete permission.”
Terminating cannot be undone. When the cluster’s persistent volumes will be deleted, the dialog lists them under Persistent volumes and asks you to type DELETE VOLUME DATA as well as the cluster’s name. Volumes kept by the provider’s retention policy keep billing at the provider after the cluster is gone.
To keep the cluster but stop paying for its servers, stop it instead on Stop, start and schedules.
In a script, --yes never stands in for --accept-volume-data-loss; without it the command stops. See Operate an Ankra Managed Cluster and, for managed Kubernetes, Managed Kubernetes.

Next steps