Patch Management Policy

Emergency patch policy

An emergency clause says when normal patch windows are suspended, who decides and how it ends. CISA's Known Exploited Vulnerabilities catalog held 1,786 entries on 18 November 2026, 302 added in 2026, and Mandiant measured an average 5 days from disclosure to exploitation in 2023.

What the section must say

  • The triggers that allow an emergency to be declared.
  • Who declares, and the deputy.
  • Exactly what is suspended (windows, lead times, approvals) and for which assets.
  • What is recorded at declaration and at closure.
  • How the emergency ends and the standard windows resume.

This page covers the policy clause only. Running a response to a new disclosure is a different subject and is outside this site.

What public text supports

NIST SP 800-40 Rev. 4 defines emergency patching as the procedure to address patching emergencies in a crisis situation, such as a severe vulnerability or one being actively exploited, and recommends the same general approach as routine patching on a highly accelerated schedule: a canary period of a few minutes to a few hours, then patching in the following hours or days depending on urgency. It separately names emergency mitigation, used to temporarily mitigate vulnerabilities before a patch is available, and says organizations should plan to replace emergency mitigations with permanent fixes on enforced schedules. DORA's RTS Article 10(4) asks for emergency patching procedures.

Counts are from the CISA feed and the Mandiant analysis published 15 October 2024; time-to-exploit figures are Mandiant's and cover that sample only.
Trigger candidateEvidence a clause can citeRead 19 Nov 2026
Listed in the CISA KEV catalog1,786 entries, version 2026.11.18; 302 added in 2026 through 18 Nov (the site's count from the feed); 361 flagged as known ransomware campaign usePrimary, JSON feed
Vendor emergency or out-of-band updateMicrosoft lists out-of-band updates as resolving a recently identified issue or vulnerability, as neededVendor
Speed of exploitationMandiant: average time to exploit 5 days for 2023 (32 days across 2021 and 2022), 138 vulnerabilities analyzed, 15 outliers excludedThird-party

Sample wording

7.1 The policy owner, or the deputy [role], may declare an emergency for named assets when (a) a vulnerability affecting them is listed in the CISA Known Exploited Vulnerabilities catalog, (b) the vendor issues an emergency or out-of-band update for software they run, or (c) the policy owner judges that waiting for the standard window is unacceptable.

7.2 A declaration records: who declared, date and time, the assets, the reason and the expected end.

7.3 During an emergency the standard windows and change lead times are suspended for the named assets only. All other assets keep their windows.

7.4 A canary installation precedes the rollout, lasting minutes to hours. Where no patch exists, the mitigation in Annex A applies and is replaced by the patch on a scheduled date.

7.5 The emergency closes when the patch is installed and verified on all named assets. The policy owner records the closure and the standard windows resume.

Common mistakes

  • Global suspension. A clause that suspends all windows invites skipping tests everywhere; scope it to named assets, as 7.3 does.
  • No end. NIST wants mitigations replaced by permanent fixes on enforced schedules; clause 7.4 and 7.5 do that.
  • Declaring by committee. One role and a deputy can act in minutes, not days.

Link the clause to severity and patch windows for class 1, to testing and staging for the canary, and to the general policy for the surrounding text.

Sources

Last reviewed 19 Nov 2026