Patch Management Policy

NIST SP 800-40 Rev. 4 patch management

NIST SP 800-40 Revision 4 (April 2022) gives no fixed number of days for any patch. It asks each organization to sort assets into maintenance groups and write a maintenance plan, with its own timeframes, for four scenarios. It is guidance, not a mandate. Read in full on 19 Nov 2026.

Why this publication sets no number

The NIST publication page defines enterprise patch management as "the process of identifying, prioritizing, acquiring, installing, and verifying the installation of patches, updates, and upgrades throughout an organization." Section 3.5 then leaves the clock to the reader: a maintenance plan defines the actions for a scenario, "including the timeframes for beginning and ending each action." The only day count in the text is not NIST's. It reports that CISA's BOD 22-01 required federal agencies to fix new KEV entries within two weeks, a rule since replaced by BOD 26-04.

The document frames the trade-off a policy window has to settle. Deploying patches faster "reduces the window of opportunity for attackers but increases the risk of operational disruption because of the lack of testing"; testing first does the reverse.

Four scenarios, each with its own plan

NIST SP 800-40 Rev. 4, Sections 3.3 and 3.5.1 to 3.5.4. Quotations are verbatim; the third column summarizes the section.
ScenarioWhat NIST says it isWhat the plan should contain
1. Routine patching"the standard procedure for patches that are on a regular release cycle and have not been elevated to emergency status"Phased deployment with canary assets; flexibility on when, then forced installation after a grace period
2. Emergency patching"the procedure to address patching emergencies in a crisis situation"The same approach on an accelerated schedule; canary period of a few minutes to a few hours; patching within hours or days
3. Emergency mitigation"to temporarily mitigate vulnerabilities before a patch is available"Quick mitigations such as deactivating functionality or isolating an asset; a schedule to replace them with the patch and remove the mitigation
4. Unpatchable assets"implementation of isolation or other methods to mitigate the risk of systems that cannot be easily patched"An approved set of long-term methods per group, reviewed in advance, and periodic re-evaluation of alternatives to patching

The document warns that routine patching "is often postponed and neglected" and that delay "makes emergency patching more difficult, time-consuming, and disruptive because of the need to first install previous patches that new patches depend upon."

Maintenance groups instead of exception lists

A maintenance group is "a set of assets with similar characteristics that generally have the same software maintenance needs for each risk response scenario." NIST's advice is to build groups rather than exceptions: "Instead of denoting certain assets or types of assets as 'exceptions,' there should be maintenance groups for them." Its examples give each group the software to patch, outage restrictions, existing mitigations and impact if compromised.

NIST SP 800-40 Rev. 4, Section 3.4 (four of six examples shown).
Example group in the textOutage restrictionImpact if compromised
Mobile workforce laptopsTolerant to downtimeModerate
On-premises datacenterScheduled outage windows for all non-emergency situationsHigh
Legacy OT assetsScheduled outage windows; software cannot be patchedHigh
Containers with customer-facing applications in the public cloudHighly tolerant to downtimeHigh

Exceptions, grace periods and metrics

Some exceptions are inevitable, so NIST asks organizations to "closely track and monitor all exceptions to maintenance plans" and to review them "regularly to determine if the maintenance plan can be implemented now." Assets with similar long-term exceptions move into a group of their own. On routine patches the text says to "offer flexibility with how soon routine patches are to be installed, while also forcing installation after a grace period has ended," either by triggering installation or by cutting network access for unpatched assets.

Section 3.6 rejects a single percentage. "Overly simplistic metrics, such as counting the number of vulnerabilities that the entire organization had and what percentage of them were patched, are not actionable." Its notional matrix reports, per cell of asset importance and vulnerability importance, the share patched by deadline plus mean and median days; the example's critical-vulnerability, high-importance cell reads 95.2 % by deadline with a 5.2 day mean. The numbers are illustrative, not a benchmark.

What a policy must write

  • A list of maintenance groups and the criteria for membership, with a named group for assets that cannot be patched.
  • One plan per group per scenario, each with start and end timeframes (the days are the organization's choice).
  • The process that decides which plan applies and when to switch plans as understanding of risk changes.
  • A canary or phased-deployment rule and the length of the grace period before forced installation.
  • An exception review interval and the rule for moving long-term exceptions into their own group.
  • Metrics that weigh asset and vulnerability importance, not one overall percentage.

The day counts that fill these plans belong in the severity and patch window section. Scenario 4 is worked through in patching systems that cannot be rebooted, and the exceptions step in patch exceptions clause.

Evidence a reviewer can ask for

The text names no auditor. What it makes checkable is the plan itself, the group membership list, the record of deadlines met per cell of the metrics matrix, and the dated exception reviews. For assets where patch history is tracked, the document notes the history "may be particularly helpful to incident responders during an investigation."

Questions

patch management policy nist

NIST SP 800-40 Rev. 4 is the usual reference. It asks for maintenance groups and a maintenance plan per scenario, and it sets no fixed day count. Each plan carries the organization's own start and end timeframes, and exceptions are reviewed regularly.

Sources

Last reviewed 19 Nov 2026