kubectl escalate status
kubectl escalate status [flags]
List active kube-escalate ClusterRoleBinding and RoleBinding objects.
Without --all, only the current user's escalations are shown.
Flags
| Flag | Default | Description |
|---|---|---|
--all | false | Show escalations for all users |
--kubeconfig string | $KUBECONFIG | Path to kubeconfig |
Output
REQUESTER ROLE SCOPE EXPIRES AT REMAINING
eike@layer87.de cluster-admin cluster 2026-05-24T16:00:00Z 55m0s
eike@layer87.de editor ns/tenant-acme 2026-05-24T15:30:00Z 25m0s
| Column | Description |
|---|---|
REQUESTER | Username from kube-escalate/requester annotation |
ROLE | Bound ClusterRole or Role name |
SCOPE | cluster for ClusterRoleBinding; ns/<namespace> for RoleBinding |
EXPIRES AT | RFC3339 expiry timestamp, as originally requested |
REMAINING | Time until expiry, or EXPIRED if past the deadline |
EXPIRES AT may not be the real expiry
status reads the kube-escalate/expires-at annotation exactly as written
by the plugin. If a request exceeded the operator's maxDuration ceiling,
the operator enforces a shorter, clamped expiry but never rewrites this
annotation (see Architecture),
so EXPIRES AT/REMAINING can overstate how long the binding actually
lives. Watch for an EscalationClamped Kubernetes Event for the real
effective expiry.
Examples
# Your own escalations
kubectl escalate status
# All users (requires list on clusterrolebindings)
kubectl escalate status --all