Kubernetes inspections and resources

Before you begin

• The cluster belongs to the current tenant and project. Its read connection and credentials are valid and cover the namespaces you intend to inspect.
• Resource reads and operations can require different permissions. Do not grant full cluster administration merely to enable log viewing.
• Check the inspection time and scope. Partial results, unreachable clusters and expired credentials must not be interpreted as a healthy empty inventory.

Move from findings to resources

1. Open the target cluster’s inspection page and check the run time, inspected resources and findings.
2. Select a finding, read its rationale and open the referenced resource. Verify cluster, namespace, type and name.
3. Review status, related resources and events. For a Pod, also inspect container state and logs rather than relying only on a finding’s title.
4. If action is needed, open the relevant operation and verify its preselected target. Opening the page does not execute it; follow actual permission and approval requirements.

Filters and result scope

Narrow the inventory by resource type and namespace, then use labels to locate objects. Label options come from returned resources; filtering a limited result does not search the whole cluster.

If only part of the inventory is returned, narrow the namespace or resource-type scope further. Screenshot counts and current-page rows are not automatically the full cluster size.

Inspections versus AI analysis

Automated inspections use rules and can produce findings without AI. AI-assisted analysis helps interpret information and propose responses. Neither replaces checking the actual resource. After a change, read the resource again or rerun the check before concluding it recovered.

Common problems