FedRAMP 20x Vulnerability Sampling: Building a Defensible Coverage Plan
Sampling effectively identical resources is permitted unless it reduces detection efficiency or effectiveness. Record why the sample represents the group and revisit that decision after changes. Keep recommended sample detection separate from mandatory machine verification and validation, even where their timeframes match.
In this article
Main question
How can Class B and C teams build a defensible FedRAMP 20x vulnerability sampling plan?
FedRAMP 20x vulnerability sampling starts with a coverage question: why does the selected resource represent the others? A fleet sharing a deployment name can still contain different software, configuration or exposure. For Class B and C teams, a useful plan explains the grouping, records the evidence and reconsiders the group when those assumptions change.
The plan below is an implementation approach. The official rules define when sampling is permitted and distinguish recommended sample detection from mandatory machine verification and validation.
Preserve the condition attached to sampling
VDR-CSO-SIR permits sampling effectively identical information resources, especially machine-based resources, for vulnerability detection unless doing so would decrease detection efficiency or effectiveness. VDR-CSO-DAC separately recommends automatic vulnerability detection on representative samples of new or significantly changed resources. These are MAY and SHOULD provisions respectively. FedRAMP Vulnerability Detection and Response
Do not turn the permission into a fleet-wide assumption. As an engineering practice, write down the characteristics used to establish effective identity: deployed artifact, relevant configuration, enabled components and exposure conditions. Decide how drift would be detected and which differences would cause a resource to leave the group. These suggested criteria are not an official percentage or sample-size formula.
For example, replicas deployed from one image may look like one group. A replica with an additional plugin or different network exposure deserves a fresh grouping decision. The practical question is whether the selected sample still represents what the detection process needs to find.
Separate three decisions in the coverage plan
Keep the permission to sample, the detection cadence and the verification obligation visible as separate decisions. For the Class B and C scope addressed here, the official page specifies:
| Provision | Class B | Class C |
|---|---|---|
| VDR-TFR-PSD: persistent detection on representative samples of similar machine-based resources | SHOULD, at least every 7 days | SHOULD, at least every 3 days |
| VDR-TFR-MVX: verify and validate machine-based resource status | MUST, at least every 7 days | MUST, at least every 3 days |
Matching timeframes do not make these the same activity or give both rules the same force. The recommended-rule decision guide explains how to handle the documentation implications of recommendations.
The VDR page lists optional adoption from July 4, 2026, initial and ongoing certification adoption on December 7, 2026, and a grace-period end of March 7, 2027. Plan against these dates and the applicable class, rather than treating a page-format update as a new sampling policy.
Make the coverage decision inspectable
A compact working record can contain the following editorial fields:
- Group identity and the inventory snapshot used to define membership.
- Reasons the resources are effectively identical for detection purposes.
- Selected resources, selection rationale and detection results.
- Evidence that sampling does not reduce detection efficiency or effectiveness.
- Changes or failures that trigger regrouping or broader detection.
- Owner and date of the most recent coverage review.
Link the record to the actual inventory and results instead of copying a stale resource count into a document. Retain enough context to explain a historical result after the fleet changes.
Exercise changes and exceptions
Try a controlled change to a resource characteristic used by the grouping logic. Check whether the process notices the difference, re-evaluates membership and schedules the appropriate detection work. Then test a failed or incomplete detection run: an attempted scan should not silently become a successful coverage result in your own reporting.
Use the evidence readiness checklist to connect these decisions to retrievable evidence. A defensible coverage plan lets another engineer reproduce why the group was selected, what was checked and when the assumptions stopped holding.
Frequently asked questions
Does VDR-CSO-SIR prescribe a fixed sample percentage?
The rule permits sampling effectively identical resources subject to the efficiency and effectiveness condition. The working record in this article is an implementation suggestion, not an official sample-size formula.
Is the Class C three-day sample-detection cadence a MUST?
VDR-TFR-PSD uses SHOULD for sample detection at least every three days. VDR-TFR-MVX separately uses MUST for Class C machine-based resource verification and validation at least every three days.
When does the current VDR page list adoption?
It lists December 7, 2026 for initial and ongoing certification adoption, with a grace-period end of March 7, 2027; optional adoption starts July 4, 2026.
Next step
If you want to turn this guidance into an execution plan, the product side handles control mapping, SSP drafting, and evidence collection.
Related articles
FedRAMP 20x: Do You Need a Separate Government Deployment?
Compare shared and dedicated deployment designs for Class B/C using the shared-infrastructure policy, actual assessment scope and agency needs.
FedRAMP 20x: Handle Denied Agency Package Access Requests
Handle denied agency package-access requests for Class B/C providers using a compatible trust center, preserving the five-business-day notification trigger.
FedRAMP 20x: Reconcile Agency Access Records in Your Trust Center
Reconcile Class B/C trust-center permission history and access activity, with the right six-month summary retention and request-specific retrieval.