Lab 5: Backup Resilience
Why it matters: A backup only helps if you can recover from it. Ransomware crews target backups first, so encryption, deletion resistance (immutability), and practiced restores are core resilience controls.
What you’ll practice
- Treating restore tests as evidence, not assumptions
- Checking encryption and immutability claims
- Reasoning about evidence freshness
Run it
From the root of the CyberSHIELD Labs repository:
python3 -m labs.backup_resilience.validate
backup_status.json lists two simulated backups. customer-records-demo is encrypted but not immutable, and its last restore test was in October 2025. analytics-demo meets all checks (as of November 2026).
Example output:
$ python3 -m labs.backup_resilience.validate
[REVIEW] customer-records-demo: immutability_not_confirmed
[REVIEW] customer-records-demo: restore_test_stale
What the checks look for
| Check | Triggers when |
|---|---|
encryption_not_confirmed | encrypted is not true. |
immutability_not_confirmed | immutable is not true. |
restore_test_missing | No restore-test date is recorded. |
restore_test_stale | The last restore test is more than 90 days old. |
restore_test_in_future | The restore-test date is in the future, which suggests bad evidence. |
Try this
- Fix
customer-records-demoin a copy of the fixture and rerun until it’s clean. - Set a restore date in the future. Why is impossible evidence worth flagging?
- Decide what restore-test interval fits a system you know, and what recovery time objective it should meet.
What this lab doesn’t do
The 90-day threshold is a teaching default, not a universal policy. The lab reads self-reported JSON and can’t verify that a backup or restore actually worked.
Takeaway
Test restores, measure recovery objectives, and protect backup administration separately from production access.
Source: labs/backup_resilience on GitHub.