Skip to main content
DELETE
Delete node group (browser session)

Authorizations

ankra_session
string
cookie
required

Browser session. Mutations also require X-Ankra-CSRF.

Headers

X-Ankra-CSRF
string
required

Must match the ankra_csrf browser cookie.

Path Parameters

cluster_id
string<uuid>
required
group_name
string
required

Query Parameters

force_drain
boolean
default:false

Take every worker of the group down on a cluster teardown's drain, all at once, without honouring PodDisruptionBudgets, and drop the group's settings with the request. Off by default: the workers are drained one at a time through their budgets, a worker whose drain is refused is kept together with its node group and the group's settings, and the operation fails naming what refused it; the settings go when the group's last worker has.

Response

Successful response

deleted
integer
required
dispatch
enum<string>
required

What the write did about the operation that applies it. started: operation_id names it, created in the same transaction as the write. deferred: the change is stored but no operation was started for it now; dispatch_reason says why and what applies it instead (the reconcile loop once a running operation has finished, or the cluster's next start). none: the write left nothing for an operation to apply (a scale to the group's current count, a removal on a stopped cluster, which the write applies to the group's records itself).

Available options:
started,
deferred,
none
group_name
string
required
operation_id
string<uuid> | null
required

The operation that applies this change, created in the same transaction as the write: poll it on /api/v1/org/executions/{execution_id} to know when the nodes have moved. Null when the write started none; dispatch says why.

dispatch_reason
string

Why no operation was started (deferred, none), or what a started operation leaves to the reconcile loop (an offline cluster's agent work). Absent when a started operation carries the whole change.