NIST SP 800-53 Rev. 5 · Contingency Planning
CP-9 on AWS
Backup missing for assets requiring it — a published SQL (Athena) + Lambda query you can run against Amazon Athena over AWS Backup job history + supplied inventory, with the evidence it needs, the settings you decide, and what a result does and does not establish.
Why this matters
A critical system whose backups have silently been failing
Backups are usually configured once and then trusted. A backup job that begins failing does so quietly, and the failure is typically discovered at the worst possible moment — during an attempted recovery after ransomware or hardware loss.
Why this check earns its place. It checks that backups actually succeeded recently, rather than that a backup was configured at some point.
- What it reads
- The organization's own backup job history, compared against its inventory of systems that require backup.
- What it reports
- Systems that require a backup and have no recent successful one.
See related public disclosures on the Signals page →
What this finds
Asset declared as requiring backup has no successful backup within twice its recovery point objective.
The requirement
The organization must conduct backups of user-level, system-level, and system documentation information at an organization-defined frequency, and protect the confidentiality, integrity, and availability of backup information.
NIST SP 800-53 Rev. 5 — CP-9 (System Backup).
NIST requires backups at a defined frequency. Which assets require backup, the recovery point objective per asset, and what counts as a successful backup are organization-defined — supplied through the asset inventory.
How this maps to the CISA performance goals
The Cross-Sector Cybersecurity Performance Goals are a voluntary baseline published by CISA — the U.S. federal agency for critical-infrastructure cybersecurity — written for organizations too small to employ security staff. This control speaks to:
- 3.O — Maintain system backups & restoration ability
The goal is maintaining backups; the pack finds assets that require one and do not have it.
See the full coverage map, including the goals no query can answer →
Evidence this query needs
Asset inventory asset_inventory.csv
Declares which assets are in scope and what is required of them. A backup that never ran on an asset you did not know existed produces no alert — absence of evidence is invisible without an inventory to compare against.
Minimum fields: asset_id, environment, criticality
Three fields let you scope the checks. The backup_required and logging_required flags are what make the absence checks possible — without them the tool cannot tell an asset that should be backed up from one that should not.
The query
Published as cp-9-backup-coverage.aws.sql. Reads aws_backup_jobs, asset_inventory (supplied). Requires: backup:ListBackupJobs; athena:StartQueryExecution.
Before this returns anything,
aws_backup_jobs has to be reaching Amazon Athena over AWS Backup job history + supplied inventory, and asset_inventory (supplied) has to be landed there. If not, the query reports nothing found — which looks
exactly like nothing wrong. What has to be switched on first →
-- CP-9 -- assets requiring backup with no recent successful job
WITH last_good AS (
SELECT
lower(resource_arn) AS asset_key,
max(completion_date) AS last_good_backup
FROM aws_backup_jobs
WHERE state = 'COMPLETED'
AND completion_date > current_timestamp - interval '7' day
GROUP BY lower(resource_arn)
)
SELECT
a.asset_id, a.asset_name, a.criticality, a.owner,
coalesce(a.rpo_hours, 24) AS rpo_hours,
g.last_good_backup,
date_diff('hour', g.last_good_backup, current_timestamp)
AS hours_since_backup,
'AWS Backup' AS source_system
FROM asset_inventory a
LEFT JOIN last_good g
ON lower(a.asset_id) = g.asset_key
WHERE a.backup_required = true
AND (a.decommission_date IS NULL OR a.decommission_date > current_date)
AND (
g.last_good_backup IS NULL
OR date_diff('hour', g.last_good_backup, current_timestamp)
> coalesce(a.rpo_hours, 24) * 2
)
ORDER BY hours_since_backup DESC NULLS FIRST;Open this pack on AWS to adjust the settings and download it →
The interactive version opens on AWS — the platform on this page — even if your environment profile does not list it.
Settings you decide
NIST states the objective and leaves these to your organization. The interactive version of this pack applies them to the query for you.
- RPO breach multiplier (default 2)
- An asset is reported once the time since its last successful backup exceeds its RPO multiplied by this factor.
- Limit to specific assets (optional)
- Paste asset identifiers, one per line.
Running it continuously
EventBridge Scheduler → Lambda → Security Hub
Schedule: EventBridge Scheduler, rate(6 hours)
Compute: Lambda (Python 3.12), 180s timeout
Step 1: ListBackupJobs for the window, filter COMPLETED
Step 2: Left-join the supplied asset inventory
Step 3: Apply per-asset RPO thresholds
Step 4: BatchImportFindings -> Security Hub
Severity label: HIGH
RelatedRequirements: NIST.800-53.r5 CP-9
Idempotency: Id = sha256(asset_id + "backup-gap")NULLS FIRST in the ordering puts assets with no backup at all above assets with a merely stale one — those are the more serious cases and should be at the top of the queue.
What a result does not prove
Absence of a backup job record does not establish that data is unrecoverable. It establishes that recoverability cannot be evidenced from the queried source — which for a resilience control is the question that matters at review.
Validation status for this platform: Not tenant-validated. Structure, query generation, and preflight are covered by the project's automated test suites. This pack has not been executed against a live tenant, account, or index. Validate it in a non-production scope before relying on any result.
Before you act on a result
- Is the asset genuinely in scope, or is the inventory flag stale?
- Is the asset protected by a mechanism this query does not read (snapshots, replication)?
- Has the backup job been failing silently, or was it never configured?
- When was the last successful restore test for this asset?
Common questions
- How do I check for backup missing for assets requiring it in AWS?
- Run the published SQL (Athena) + Lambda query on this page against Amazon Athena over AWS Backup job history + supplied inventory. It reads aws_backup_jobs, asset_inventory (supplied) and reports: Asset declared as requiring backup has no successful backup within twice its recovery point objective.
- What permissions are needed to run this AWS query?
- backup:ListBackupJobs; athena:StartQueryExecution.
- What evidence does CP-9 need that AWS cannot produce?
- Asset inventory (asset_inventory.csv). Declares which assets are in scope and what is required of them. A backup that never ran on an asset you did not know existed produces no alert — absence of evidence is invisible without an inventory to compare against. At minimum it must carry: asset_id, environment, criticality.
- Does a result from this query mean the control has failed?
- Absence of a backup job record does not establish that data is unrecoverable. It establishes that recoverability cannot be evidenced from the queried source — which for a resilience control is the question that matters at review.
Terms used on this page
- NIST SP 800-53the U.S. federal catalogue of security requirements
- The catalogue of security and privacy requirements published by the National Institute of Standards and Technology, a U.S. federal agency. U.S. government systems are measured against it, and many private organizations adopt it voluntarily. Each requirement has an identifier such as AC-2. Official source →
- CISACybersecurity and Infrastructure Security Agency
- The U.S. federal agency responsible for national critical-infrastructure cybersecurity. Its publications are works of the U.S. government and are free to use. Official source →
- CPGCross-Sector Cybersecurity Performance Goals
- A voluntary baseline of security practices published by the U.S. Cybersecurity and Infrastructure Security Agency, written specifically for organizations too small to employ security staff. It is a short, plainly written starting list rather than a full standard. Official source →
- RPOrecovery point objective
- How much recent work an organization has decided it can afford to lose. A four-hour objective means backups must be recent enough that no more than four hours of work would be lost.
- not tenant-validatednever executed against real data
- The check has passed the project's automated tests for structure and syntax, but has never been run against real data anywhere. Test it in a safe scope before relying on any result.