07 principle
Rollback Without Fear
When something goes wrong — and it will — the recovery path should be as structured as the deployment path. Rolling back a DNS change should never involve guesswork, manual console edits, or hoping you remember what the record used to be.
A deployment rollback should show the operator exactly what will happen before it happens:
- Which items will be reversed (and how — delete, restore, or recreate)
- Which records were modified by later deployments (cascading risk)
- Which records were changed outside the tool (external drift)
- Which items can’t be rolled back (insufficient state captured)
Confirm, and every item is reversed in the correct order with a full audit trail. Individual record rollback should also be available from the audit log for surgical corrections.