Internal SLAs for remediation and exceptions: deadlines that can be proved
A remediation deadline needs a clock start, a breach rule and a route to an exception.

An internal SLA (service-level agreement) for remediation is a governance instrument, not a promise: to be provable it needs an owner, a clock start, pause conditions, a breach consequence and a route into the exception path. And no NIST publication prescribes an internal remediation SLA. SP 800-53 Rev. 5 makes the organisation set the time period (SI-2) and measure it (SI-2(3)); the plan of action and milestones (POA&M) gives the milestone-and-date mechanics; the one prescribed clock in this corpus belongs to CISA’s Binding Operational Directive 26-04, a documented external regime for US federal agencies, not universal advice.
An internal SLA is a governance instrument, not a promise
Five parts, and a deadline missing one is a date nobody can administer: an owner accountable for the deadline, not only the work; a clock start, the event that begins the count; pause conditions, the changes of fact that shift the class and who may declare them; a breach consequence, what changes when the date passes; an exception route, a missed deadline turned into a recorded, time-bound deviation. Vocabulary: SP 800-100 uses “service-level agreement” for the document carrying “appropriate actions and maximum response times” in incident prioritisation (lines 6913–6914), not for remediation deadlines — calling an internal remediation clock an SLA is our reading of a pattern the sources do not name.
What NIST requires: a time period you set, then a measurement
The catalogue states the duty and leaves the number to you. SI-2(c) requires updates installed “within [Assignment: organization-defined time period] of the release of the updates” (line 21749): the deadline is mandatory, its value is the organisation’s. SI-2(3) closes the loop — measure the time between flaw identification and remediation and set organisation-defined benchmarks (lines 21791–21794) — benchmarks that “can be established by the type of flaw or the severity of the potential vulnerability” (lines 21803–21804). SP 800-40r4 puts the deadline in the maintenance plan, which defines a scenario’s actions “including the timeframes for beginning and ending each action” (lines 893–894). Verified negative: SP 800-53r5 contains no “within N days” prescription, and no NIST document in this corpus supplies one.
The deadline needs a milestone and a date
A deadline that is not a dated milestone on a tracked plan is not a deadline anybody owns. CA-5 requires a plan of action and milestones “to document the planned remediation actions” for weaknesses found in assessment (lines 6636–6639). SP 800-37r2 shapes it with “milestones established to meet the tasks; and the scheduled completion dates for the milestones and tasks” (lines 4634–4636), and ties the same plan to corrective action — “a cost-effective method for initiating corrective actions” (line 4110) — and to the time-based triggers of ongoing authorisation (line 3870). NISTIR 8286r1’s POA&M box repeats the pair (lines 1828–1830). RA-7 supplies the trigger: a plan of action and milestones entry is generated when a risk response is to mitigate a risk and the mitigation “cannot be completed immediately” (lines 16273–16274).
A documented external regime: the clock start and the calendar
CISA’s BOD 26-04 is the corpus’s worked specification of a prescribed clock; it binds US federal civilian executive branch agencies only. It assigns each vulnerability a timeline from four variables — asset exposure, KEV (Known Exploited Vulnerabilities) status, exploit automation, technical impact (lines 97–101) — and the clock starts with whichever happens first: CISA adding the vulnerability to the KEV Catalog, or the agency identifying it on an asset, since “whichever event occurs first starts the remediation timeline” (line 181). Days are calendar days (line 176), not working days.
The clock that moves, and the rules that bend it
A deadline under this regime is not fixed when it is assigned: “The timelines for each asset are dynamic as facts change”, so removing a system from the internet shifts the deadline back and a KEV entry shortens it (line 180). Two rules show how specific a class can be — a “+ forensic triage” row means remediation within three days plus a triage of the asset (line 179), and “fix on system upgrade” defers the fix to the next scheduled rebuild (line 182).
clock start class the asset falls into outcome
| | |
identification -- first of --> exposure x KEV x exploit x impact --> deadline (calendar days)
or external listing | |
met <-----+ +----> missed
(evidence: the condition (exception entry:
is gone) owner, end date)
^ the clock moves when a fact moves: exposure drops -> deadline shifts back;
the vulnerability enters the KEV -> deadline shortens
Breach: what a missed date changes
Enforcement separates a deadline from a target. SP 800-40r4 tells organisations to “offer flexibility … while also forcing installation after a grace period has ended” (line 940), and requires that “Schedules should be set and enforced for both patch deployment and mitigation removal” (lines 964–965). A deadline with no consequence is a report, not an SLA.
The pause, and the road into the exception path
A pause needs the same discipline as a breach. The dynamic-timeline rule is the source’s own step rule, so an internal policy must name the fact that reclassifies the asset and who may reclassify it. The road then ends at the exception path, not at an extension by habit: SP 800-40r4 asks organisations to “closely track and monitor all exceptions to maintenance plans” and review them “regularly” (lines 995–998), and CSF 2.0’s ID.RA-07 requires that “Changes and exceptions are managed, assessed for risk impact, recorded, and tracked” (lines 905–906). The approval machinery is sheet 1; the row is the L2 register’s.
Proving the deadline: the numbers to report
A deadline is provable when the organisation reports two things per class. SP 800-40r4’s notional metrics “reflect the percentage of assets that were patched by the corresponding maintenance plans’ deadlines” (lines 1054–1055), alongside average and median time to patch and target values for improving them (line 1082). Reading, labelled as ours: the first number is the share closed inside the deadline, the second the share that ended in an exception instead of a fix. The mean hides the tail; the by-deadline percentage and the exception rate are what a governing body can hold someone to, and GV.RM-04’s strategic direction on risk response options (line 772) plus GV.RM-06’s standardised documenting method (lines 778–779) make them comparable period to period. ID.RA-06 frames the duty: “Risk responses are chosen, prioritized, planned, tracked, and communicated” (line 903).
Level and prerequisites
L3 — assessment, control selection and process coordination: choosing the clock, the classes, the pause conditions and the breach consequence, and proving the result. It assumes the L2 register rows and the L1 vocabulary of acceptance; the severity reading is sheet 5, the priority decision sheet 6.
Where to go next
- Knowledge — the index.
- Cybersecurity Governance — the area this sheet belongs to.
- Risk management — this sheet’s node.
Related sheets are unpublished and named in prose only: the L2 risk and exception register, prioritisation (sheet 6), the exception process (sheet 1) and verification (sheet 8).
References
- NIST — SP 800-53 Rev. 5, Security and Privacy Controls for Information Systems and Organizations — https://doi.org/10.6028/NIST.SP.800-53r5 — SI-2(c) (the organisation-defined time period), SI-2(3) (measure time to remediate and set benchmarks), CA-5 (the plan of action and milestones), RA-7 (the POA&M entry when a mitigation cannot be completed immediately).
- NIST — SP 800-40 Rev. 4, Guide to Enterprise Patch Management Planning: Preventive Maintenance for Technology (2022) — https://doi.org/10.6028/NIST.SP.800-40r4 — §3.5 (maintenance plans that carry timeframes; the grace-period enforcement; exceptions to maintenance plans) and §3.6 (the by-deadline and time-to-patch metrics, with target values).
- NIST — SP 800-37 Rev. 2, Risk Management Framework for Information Systems and Organizations — https://doi.org/10.6028/NIST.SP.800-37r2 — Task A-6 and the POA&M discussion (tasks, milestones and the scheduled completion dates for both), the corrective-action framing and the time-based trigger of ongoing authorisation.
- NIST — NISTIR 8286r1, Integrating Cybersecurity and Enterprise Risk Management (ERM) (December 2025) — https://doi.org/10.6028/NIST.IR.8286r1 — the POA&M box: milestones in meeting the tasks and scheduled completion dates for the milestones.
- NIST — The NIST Cybersecurity Framework (CSF) 2.0, NIST CSWP 29 — https://doi.org/10.6028/NIST.CSWP.29 — ID.RA-07 (exceptions recorded and tracked), ID.RA-06 (responses planned and tracked), GV.RM-04 (strategic direction on risk response options), GV.RM-06 (a standardised documenting method).
- NIST — SP 800-100, Information Security Handbook: A Guide for Managers (October 2006) — https://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication800-100.pdf (no DOI printed in this edition) — used only for its single use of “service-level agreement” as the document carrying maximum response times in incident prioritisation.
- CISA — Binding Operational Directive 26-04: Prioritizing Security Updates Based on Risk (10 June 2026) — https://www.cisa.gov/news-events/directives/bod-26-04-prioritizing-security-updates-based-risk — the four decision variables, Table 1 and its calendar-day timelines, the clock-start rule, the dynamic-timeline rule, the forensic-triage and fix-on-upgrade rules, the seven-day reporting duty and the annual reassessment clause.
- CISA — Known Exploited Vulnerabilities Catalog — https://www.cisa.gov/known-exploited-vulnerabilities-catalog — the catalog is an input to an organisation’s prioritisation framework, not the framework.
- CISA — Binding Operational Directive 22-01 (REVOKED) — https://www.cisa.gov/news-events/directives/bod-22-01-reducing-significant-risk-known-exploited-vulnerabilities-revoked — cited only as the historical regime superseded by BOD 26-04; its two-week remediation rule is not current practice.