Manage Machine Groups

Machine groups provide the virtual machines on which your applications run. Open the cluster and select Manage machine groups to see each group's status, Kubernetes image version, upgrade availability, machine count, resource template, and disk size.

Select Add machine group, then configure:

• Name: a unique name within the cluster.
• Image: the operating-system and Kubernetes machine image provided for the cluster version.
• Template: the CPU and memory allocated to each machine.
• Root disk size: the system disk for each machine; it must be within the allowed range and a multiple of 5 GB.
• Number of machines: a fixed number, or a minimum and maximum when autoscaling is enabled.
• SSH keys: optional public keys for direct machine access. Add them only for people who need infrastructure-level access.
• Label and Taint values: optional Kubernetes scheduling settings.

The estimated cost changes with the resource template, disk, and machine count. Insufficient balance or quota can prevent creation; reduce the requested capacity or resolve the project limit and try again.

When autoscaling is disabled, Number of machines is the group's desired capacity. Increasing or decreasing it adds or removes machines.

When autoscaling is enabled, set the minimum and maximum. The current count stays within that range and changes according to workload demand. This feature depends on Cluster management by Kubit and the Cluster Autoscaler add-on. If the Overview page warns that this add-on is missing, enable it under Add-ons.

Labels are used by selectors and affinity rules. A Taint prevents Pod scheduling unless the Pod has a matching Toleration. Available effects are NoSchedule, PreferNoSchedule, and NoExecute.

Use these options only when the application manifest is prepared for them. An incorrect selector or Taint can leave Pods in Pending, and NoExecute may remove an incompatible Pod from the machine.

Open the group and select Edit machine group. You can change capacity, autoscaling range, image, resource template, root disk, SSH keys, Labels, Taints, and pause state.

Changing only the number of machines scales the existing group. Changing other machine settings causes a rolling replacement: new machines are added one by one, and old machines are drained and removed. Before confirming, ensure that workloads have enough replicas and can tolerate disruption.

If an upgrade indicator appears, select an available newer image on the edit page. After upgrading the cluster version, upgrade older groups as suggested on the Overview page.

The pause rollout option stops group reconciliation and is not a substitute for reducing the machine count to zero. Resume the group before expecting later changes to be applied.

Select Delete machine group from the group menu. The machines in that group are deleted, and the operation cannot be undone.