Open this resource#

Console route: /poddisruptionbudgets. This is a namespaced resource; choose the namespace before making a change.

  1. Select the correct cluster in the sidebar.
  2. Open PodDisruptionBudgets and narrow the list using the available namespace/search controls.
  3. Select the exact object by name. Verify its scope and identity at the top of the detail page.

For screenshots of the shared list, detail and creation controls, see the resource actions visual guide. Available actions depend on the resource and your permissions.

What to inspect#

Pod selector and the configured minAvailable or maxUnavailable constraint, plus status.

Refresh the resource when you need the latest observation. Overview fields summarize the object; YAML exposes the complete returned document. A placeholder or missing status field should not be read as a successful or zero-valued result.

Use the available features#

Open PDBs and inspect which pods the selector covers. Compare the allowed disruptions/status with current readiness. Use the normal YAML editor for an intentional change to the budget.

Describe, Clone, YAML editing and Delete are shared resource actions where the page and your permissions allow them. Read Create, edit, clone and delete resources before applying a change.

A practical investigation#

A drain blocked by disruption policy can be protecting availability. Inspect the workload and budget before weakening it; Orkiva maintenance mode for its own database is unrelated to Kubernetes PodDisruptionBudgets.

  1. Establish the current state and the symptom you want to resolve.
  2. Identify whether a controller, Helm release or Delivery project owns the object.
  3. Make the smallest authorized change through that owner when possible.
  4. Refresh and verify the resulting resource state. Inspect related objects rather than stopping at a successful save message.

Permissions and unavailable data#

Listing this kind does not automatically grant every subresource action. The console checks operation access, and Kubernetes still checks the connected identity. If this API is not served by the cluster, fix the installation/version capability rather than treating an empty or unavailable page as an authorization grant.