Most organisations rehearse detecting and containing an attack, but far fewer have a tested plan for putting the business back together afterwards and proving the attacker has gone. NIST SP 800-184, Guide for Cybersecurity Event Recovery, fills that gap: it covers recovery planning, playbooks, continuous improvement and metrics, and it works through data breach and destructive malware scenarios. This guide explains what SP 800-184 asks for, how its tactical and strategic recovery phases work, where identity and service accounts fit, and how to map it to the Recover function of NIST CSF 2.0. It is for resilience, incident response and infrastructure leads who own recovery plans. It goes deeper on recovery than the Cyber Resilience Guide and sits alongside the NIST SP 800-61 Incident Response Guide.
Key takeaways
- Recovery is planned before the event: SP 800-184 expects asset and dependency inventories, criticality ratings, named people and a playbook to exist before anything goes wrong.
- Two phases, different goals: the tactical phase restores operations by running the playbook; the strategic phase improves the organisation's security so the event is less likely and less damaging next time.
- Do not recover into a compromised environment: the guide asks for confidence about the adversary's persistence mechanisms, or their absence, before recovery begins in earnest.
- Identities are a root of trust: forgotten and service accounts and access to management infrastructure must be protected and, after an event, reset as part of recovery.
- Measure recovery: SP 800-184 suggests metrics for damage and cost, for risk assessment improvement and for recovery quality, with cautions about misuse.
What it is and who it applies to
SP 800-184 was published as Final in December 2016 by NIST, authored by Michael Bartock, Jeffrey Cichonski and Murugiah Souppaya (NIST), Matthew Smith and Gregory Witte (G2) and Karen Scarfone (Scarfone Cybersecurity). When checked on 27 October 2026 the NIST publication page showed no withdrawal or revision notice. It is guidance, not a mandate. US federal agencies use it alongside the SP 800-53 contingency planning and incident response controls, and any organisation can use it to structure its recovery capability.
The document is written around the Recover function of the original Cybersecurity Framework and maps recovery activities to CSF core components and to SP 800-53 Revision 4 controls in Appendix B. Both have since been superseded (CSF 2.0 and SP 800-53 Rev. 5), so the mappings in this guide use the current identifiers. The recovery method itself has not dated: dependencies, playbooks, exercises and metrics are as relevant now as in 2016.
| Section | Content |
|---|---|
| 2 Planning for Cyber Event Recovery | Enterprise resiliency, recovery planning prerequisites, the recovery plan (documents, procedures, initiation and termination criteria, root cause and containment), communications, sharing recovery insights |
| 3 Continuous Improvement | Validating recovery capabilities through exercises and tests; improving recovery and security capabilities from lessons learned |
| 4 Recovery Metrics | Example metrics in Table 4-1 |
| 5 Building the Playbook | Consolidated recommendations split into tactical and strategic phases |
| 6 and 7 Scenarios | Data breach recovery and destructive malware (ransomware) recovery, each with pre-conditions, tactical phase and strategic phase |
| Appendices | A: checklist of playbook elements; B: CSF and SP 800-53 Rev. 4 mapping; C: acronyms; D: references |
Structure and key requirements
Recovery planning prerequisites
- Inventories with dependencies: people, process and technology assets, including external partners and hosted services, with their dependencies. Business impact analyses, service level and operational level agreements and dependency maps all feed this, with a focus on security dependencies. The guide's example is mail depending on LDAP: restore order follows dependency.
- Criticality: documented categorisation of assets by relative importance (FIPS 199 and FIPS 200 where applicable), so limited recovery resources go to the highest mission impact first. See the Minimum Viable Company and Critical Service Recovery Order Guide.
- People and authority: named key personnel across organisational tiers, and at least two managers who may activate the recovery plan. A privacy team member should assess risk to affected individuals.
- Planning for diminished operations: operational workarounds where recovery time objectives will be missed.
What the recovery plan contains
Section 2.3.1 lists: service level agreements; authority (management contacts who can activate the plan); recovery team membership; specific recovery details and procedures with diagrams; out-of-band communications; the communication plan; off-site storage details; operational workarounds; facility recovery details; and the infrastructure, hardware and software needed for recovery. Use it as a completeness checklist for an existing plan, or start from the Disaster Recovery Plan Template.
The two phases
| Phase | Stages | What happens | Typical owner |
|---|---|---|---|
| Tactical | Initiation, execution, termination | Run the prepared playbook with input from detection, protection and response: restore from clean backups, rebuild systems, reset credentials, tighten perimeter rules, run business workarounds; exit when defined criteria are met | Recovery manager with IT operations and IR |
| Strategic | Planning and execution, metrics, recovery plan improvement | Reduce attack surface and the likelihood and impact of future events; measure; update plans and the risk assessment | CISO, risk and business owners |
The playbook should set out major phases, procedures and exit criteria for each stage, include both technical and non-technical (business process and policy) procedures, and automate procedures where possible. Invocation timing must be agreed with the incident response team, CISO and business owners so that recovery actions do not alert the adversary or destroy evidence.
Root cause and containment before recovery
Section 2.3.4 asks for basic knowledge of the adversary's objective and high confidence about persistence mechanisms, or their absence, before recovery starts in earnest, with containment isolating compromised assets from recovered ones. In practice this means a clean recovery environment, rebuilt or verified identity infrastructure, and restored systems kept separate until verified. See the Cloud Recovery and Clean Room Guide.
How the scenarios work
Sections 6 and 7 show the method applied end to end. In the data breach scenario, a fictional organisation finds that stolen credentials gave an attacker access to critical systems; network monitoring confirms exfiltration of a significant volume of personal data, and customer financial data may also have been taken. The scenario lists the pre-conditions that made an orderly recovery possible: a formal recovery process, a list of critical assets and personnel, current dependency maps, an agreed set of recovery metrics, tested tools and a communications plan. It then walks through the tactical phase (initiation, execution, termination) and the strategic phase (planning and execution, metrics, plan improvement). The destructive malware scenario follows the same structure for ransomware. The value for practitioners is the pre-condition list: if any item is missing in your organisation, the recovery will be improvised. Use both scenarios as tabletop scripts, replacing the fictional details with your own systems, and score each run against the pre-conditions.
Identity and NHI requirements
SP 800-184 names identities, including forgotten or service accounts, as root-of-trust elements alongside access controls over management infrastructure and data integrity. It also lists resetting credentials as a playbook procedure, and its data breach scenario begins with stolen credentials. Translate that into these recovery requirements:
- Directory first: recovery order should put identity providers and directory services ahead of the applications that depend on them, and they must be verified clean. See the Identity Recovery Guide.
- Credential reset scope: include service accounts, API keys, certificates, cloud access keys, OAuth grants and automation tokens, not only user passwords. A playbook step "reset credentials" without an inventory of machine credentials will miss most of them. See the Service Account Inventory Template.
- Dormant accounts: remove or disable accounts that should not exist before bringing services back. In the Colonial Pipeline ransomware attack, attackers entered with the password of a legacy VPN account that was unused but still enabled and had no MFA; the pipeline was shut down for about six days.
- Break-glass access: recovery teams need emergency administrative access that does not depend on the systems being recovered. See the Break-Glass Account Guide.
- Management plane identities: backup, hypervisor and device management admin accounts must be separate from production identities so they survive a directory compromise. See the Backup Identity Isolation Guide.
Implementation steps
- Build the dependency map. For each critical service, list systems, data, identities, network services and third parties it depends on, and derive a restore order.
- Rate criticality and set objectives. Agree recovery time and recovery point objectives with business owners, and the workaround if they are missed. Use the Business Impact Analysis Template.
- Name authority. Document at least two managers who can activate the plan, the recovery team, and out-of-band contact routes.
- Write playbooks per event type. At minimum data breach and destructive malware, matching SP 800-184 sections 6 and 7, each with initiation criteria, procedures, exit criteria and termination criteria.
- Prepare recovery assets. Clean, tested backups; golden images; offline copies of the plan; recovery infrastructure and licences; the credential reset runbook.
- Exercise. SP 800-184 suggests tabletop exercises with an adversary role, scenarios based on news events and failover tests, followed by post-exercise reviews with tracked remediation. See the Tabletop Exercise Guide.
- Run the strategic phase after every event and exercise. Separate short-term fixes (a system not patched often enough) from long-term ones that change the risk assessment, technology or process.
- Share insights. Decide in advance what recovery information may be shared with sector bodies and partners, as section 2.5 recommends.
Choosing depth
| Organisation | Practical scope |
|---|---|
| Small | One recovery plan covering both scenarios, a dependency list for the top five services, one tabletop exercise a year, and backup restore tests |
| Mid-size | Separate playbooks per scenario, documented restore order, credential reset runbook including machine identities, twice-yearly exercises |
| Large or critical service provider | Playbooks per business service, clean room recovery, automated restore, exercises with suppliers, metrics reported to the board |
Evidence and assessment
Assessors and auditors look for evidence that recovery capability exists and has been tested. Map SP 800-184's metrics (Table 4-1) to a recovery scorecard:
| Metric group | Example metrics from SP 800-184 | Evidence |
|---|---|---|
| Damage and cost | Losses from released proprietary information; legal costs; hardware, software and labour costs of recovery; business disruption in hours, days or weeks; consequential damage such as lost customer trust | Post-incident report, finance records |
| Risk assessment improvement | Recovery exercises per year; significant incidents not identified in the risk assessment; system dependencies not identified; gaps found in exercises | Exercise reports, risk register updates |
| Quality of recovery | Business functions disrupted; stakeholder satisfaction; IT services meeting uptime requirements; timely restoration from backup; recovery events meeting objectives | Service reports, restore test logs |
SP 800-184 cautions that metric collection must not slow recovery, that metrics are less useful for unusual events and that comparisons between organisations can mislead. Add identity-specific measures: percentage of machine credentials covered by the reset runbook, time to reissue critical service credentials in exercises, and number of dormant privileged accounts found during recovery.
Appendix A's checklist of playbook elements is a practical audit tool: compare each playbook against it and record gaps.
Mapping to other frameworks
| SP 800-184 element | NIST CSF 2.0 | NIST SP 800-53 Rev. 5 |
|---|---|---|
| Execute recovery plan (tactical initiation) | RC.RP-01 | CP-10, IR-4 |
| Select and prioritise recovery actions by criticality | RC.RP-02, RC.RP-04 | CP-2, PM-11 |
| Restore from clean backups, verify integrity | RC.RP-03, RC.RP-05 | CP-9, CP-10 |
| Termination criteria and documentation | RC.RP-06 | IR-4 |
| Recovery communications | RC.CO-03, RC.CO-04 | IR-4, CP-2 |
| Exercises and improvement (strategic phase) | ID.IM-02, ID.IM-03, ID.IM-04 | CP-4, IR-8 |
| Credential reset and account clean-up | PR.AA-05 | AC-2, IA-5 |
| Dependency mapping | ID.AM-03, GV.OC-04 | CP-2 |
CIS Controls v8.1 Control 11 (Data Recovery) covers the backup portion; see the CIS Control 11 Data Recovery Guide. For destructive events, the NCCoE data integrity guides add technical capability detail; see the Data Integrity Protection Against Destructive Attacks Guide.
Common pitfalls
- A disaster recovery plan treated as a cyber recovery plan: DR assumes the environment is trustworthy and only unavailable. Cyber recovery must assume backups, identities and admin tools may be compromised.
- Recovering before containment: restoring systems while the attacker still has persistence reinfects them, which is why SP 800-184 puts root cause and containment ahead of full recovery.
- Single point of authority: only one person can invoke the plan, and they are unavailable. The guide asks for at least two.
- Credential resets that skip machines: user passwords are reset but service account passwords, API keys and certificates are not, leaving the attacker's access in place.
- Unknown dependencies: an application cannot start because a forgotten directory, DNS or licensing service is down. SP 800-184 counts undiscovered dependencies as a metric for this reason.
- No strategic phase: the organisation restores service and moves on, so the same weakness causes the next event.
Practitioner checklist
- Maintain dependency maps for every critical service, including identity and third-party dependencies.
- Agree criticality ratings, recovery objectives and workarounds with business owners.
- Name at least two managers with authority to activate the recovery plan, plus out-of-band contacts.
- Write data breach and destructive malware playbooks with initiation, exit and termination criteria.
- Include machine credentials (service accounts, keys, tokens, certificates) in the credential reset runbook.
- Require containment and persistence checks before full recovery begins.
- Exercise at least annually with an adversary role, and track remediation from every debrief.
- Report recovery metrics from Table 4-1 and identity-specific measures to leadership.
- Review each playbook against the Appendix A checklist.
Standards and references
- NIST: SP 800-184, Guide for Cybersecurity Event Recovery
- NIST: SP 800-184 full text (PDF)
- NIST: Cybersecurity Framework 2.0
- NIST: SP 800-53 Rev. 5
- NIST: SP 1800-11, Data Integrity: Recovering from Ransomware and Other Destructive Events
- CIS: Control 11 Data Recovery
Related NHI Mgmt Group resources: Cyber Resilience Guide · Ransomware Response and Recovery Guide · Business Continuity and Disaster Recovery Guide · Restore Testing and Recovery Assurance Guide