Kubernetes cluster states - Raff

_{Updated August 24, 2026}

Cluster

Status	Meaning	Your move
`pending` / `deploying`	Provisioning (10–20 min)	Watch the Activity feed
`running`	Control plane up	—
`warning`	Degraded; message has detail	Usually transient; contact support if it persists
`failed`	Provisioning could not complete — infrastructure was cleaned up, nothing bills	Delete the record; try again or contact support
`deleting`	Teardown in progress	—
`delete_failed`	Teardown incomplete; nothing was removed	Click Retry delete

The boolean ready (API and dashboard) is true only when the cluster is running and no pool is mid-operation — poll that instead of parsing statuses.

Node pools

pending → provisioning → running, with scaling / deleting while a background operation runs (the pool’s status message reports the outcome, e.g. a drain blocked by a PodDisruptionBudget).

Nodes

Status	Meaning
`provisioning`	VM deploying / joining the cluster
`running`	kubelet reports Ready
`not_ready`	Joined but failing health checks
`cordoned`	Unschedulable — a drain in progress or a manual `kubectl cordon`
`deleting`	Being drained and removed

Nodes in not_ready and cordoned still bill — the VM is running; these states describe Kubernetes scheduling, not the machine.

Last modified on August 24, 2026