Wunderbox operations

Use the selected release and environment runbooks for implementation-specific commands. This page defines the operating information those procedures should preserve.

Daily readiness

Review consumer-facing health, capacity headroom, dependency state, security signal health, backup job status where applicable, and unresolved change or incident conditions. A green component check is not enough when the service outcome is degraded.

Change preflight

Before a material platform change, confirm:

• approved objective, scope, owner, and maintenance window;
• immutable artifact and automation versions;
• management identity and least-privilege authorization;
• affected consumers, dependencies, and failure domains;
• capacity needed during failure and recovery, not just steady state;
• current backup status and a verified recovery path when state is at risk;
• safe-stop criteria before destructive or irreversible actions; and
• independent platform and consumer verification.

If observed topology, version, capacity, or backup state differs from the plan, stop before the first material change.

Backup and recovery

Do not assume every resource needs the same backup or that a platform feature protects application data automatically. Classify each stateful recovery unit, assign a data owner, set retention and recovery objectives, protect backup credentials and copies, and test restore plus service verification. Record results privately and publish no target, key, or customer detail.

Degraded operation

When capacity or a dependency is impaired, protect data integrity and management access first, bound the affected failure domain, pause unrelated changes, and communicate through the authorized process. Prefer a documented degraded mode over improvised reconfiguration whose recovery path is unknown.

Retirement

Remove consumer dependencies, revoke management identities, sanitize or retain data according to policy, remove monitoring and backup jobs, release capacity, and verify that automation no longer targets the retired unit. Preserve only the evidence required by retention rules.

The Incus image deployment and RHEL image pages apply this lifecycle to review candidates. Both require license, provenance, lab, and subject-matter review before approval.

See Troubleshooting and the cross-product backup and recovery guidance.