CloudAcademia
EN

Cybersecurity · Module 08 · Incident Response Basics

Incident Response: From Detection to Lessons Learned

Lesson55 minIntermediateMarkdown

Showing raw Markdown: the renderer could not load.

## Why this matters
At 02:14 on a Wednesday, someone logs in to a company's web server over SSH with the account of a system administrator who is asleep. Twenty minutes later, a 4.8 MB file of customer records is downloaded from the admin pages. Nobody notices until a colleague reads the logs the next morning. The question is no longer "could this happen to us?" but "what do we do now, and in what order?"

Without a plan, people improvise. Someone reboots the server and destroys the evidence. Someone else emails the whole company. Nobody knows who may talk to customers, whether a law requires a report, or how long it has been going on. **Incident response** is the discipline that replaces that panic with a calm, repeatable process: prepare, detect, understand, contain, recover and learn. You don't need to be a forensic expert to follow it. You need to know the steps, keep notes, and make decisions with the facts you have.

## Learning objectives
- Explain the incident response life cycle in NIST SP 800-61r3 and how it relates to the NIST Cybersecurity Framework 2.0 functions.
- Classify and prioritise an incident by its functional impact, the data affected and the effort needed to recover.
- Analyse authentication and web server logs to build a timeline of an incident.
- Choose containment, eradication and recovery actions, and explain how to preserve evidence while doing so.
- Write a short incident report with root cause, impact and lessons learned.

## Key concepts

### Events and incidents
Think of a smoke detector. It beeps when you burn toast and when the kitchen is on fire. The beep is not the fire. Security works the same way: most signals are harmless, and a few are emergencies.

An **event** is any observable occurrence in a system or network: a user logs in, a firewall drops a packet, a file is deleted. Millions happen every day. An **incident** is an event, or a set of events, that actually or imminently harms the confidentiality, integrity or availability of information or systems, or that violates security policy or law. (NIST defines it in these terms; the wording above is shortened.) The three properties are the "CIA triad" from the Security Fundamentals module.

| Observation | Event only, or incident? |
|---|---|
| One failed SSH login for `alice` from her home IP | Event: a typo is the most likely cause |
| 23 failed logins across 8 usernames in three minutes from one address | Possible incident: needs analysis |
| A successful login for the same address a few seconds after the failures | Incident: an account is probably compromised |
| A laptop with an unencrypted customer spreadsheet is left on a train | Incident: confidentiality is at risk, even if nobody has opened the file yet |
| A disk fails and the database is down for an hour | Incident (availability), even with no attacker |

Incidents are not always attacks. A misconfigured storage bucket, an employee emailing a file to the wrong person and a natural disaster all count. The same process handles them.

> [!CHECK]
> A firewall logs 4,000 blocked connection attempts from the Internet to closed ports in one hour. What is the best description?
>
> - [ ] A confirmed incident, because attackers are involved
> - [x] An event: normal Internet background noise, worth watching but not an incident by itself
> - [ ] Not worth logging, because the firewall already blocked them
> - [ ] An incident only if the firewall is a hardware appliance
>
> Blocked probes happen constantly. An event becomes an incident when it actually or imminently harms confidentiality, integrity or availability, or breaks policy. Keeping the logs still matters, because later events may be compared against them.

### Two NIST models: the classic life cycle and SP 800-61r3
The National Institute of Standards and Technology (NIST) publishes the reference guide for incident handling, Special Publication (SP) 800-61. You will meet two versions.

**Revision 2 (2012)** describes a four-phase life cycle that most teams still use as their mental model:

1. **Preparation**: get ready before anything happens.
2. **Detection and analysis**: notice, confirm and understand.
3. **Containment, eradication and recovery**: stop it, remove it, restore service.
4. **Post-incident activity**: learn, report and improve, which feeds back into preparation.

<figure class="diagram" style="--dg-w: 1040px"><svg xmlns="http://www.w3.org/2000/svg" class="dg" viewBox="0 0 1040 260" width="1040" height="260" role="img" aria-labelledby="ir-life-cycle-title ir-life-cycle-desc" font-family="Inter, system-ui, -apple-system, Segoe UI, Roboto, sans-serif" xml:lang="en">
<title id="ir-life-cycle-title">The classic incident response life cycle</title>
<desc id="ir-life-cycle-desc">Four numbered cards joined by arrows. 1, Preparation: roles, contacts, playbooks, logging, backups and training. 2, Detection and analysis: spot events, confirm an incident, triage and prioritise. 3, Containment, eradication and recovery: stop the spread, remove the cause and restore service. 4, Post-incident activity: lessons learned, report and evidence retention. A return arrow runs from card 4 back to card 1, labelled: lessons learned improve preparation, the cycle repeats. A footnote says this is the NIST SP 800-61r2 model from 2012, now superseded but still a useful mental model.</desc>
<rect class="dg-bg" x="0.5" y="0.5" width="1039" height="259" rx="16" fill="#ffffff" stroke="#e3e3dd"/>
<rect class="dg-card" x="24" y="24" width="222.5" height="130" rx="12" fill="#ffffff" stroke="#cfcfc8"/>
<g class="dg-l1"><circle class="dg-solid" cx="46" cy="48" r="12" fill="#0369a1"/><text x="46" y="52.5" font-size="12.5" class="dg-on" fill="#ffffff" font-weight="700" text-anchor="middle">1</text></g>
<g class="dg-l1"><g class="dg-icon" transform="translate(206.5 36) scale(1.0833)" fill="none" stroke="#0369a1" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round"><path d="M8 6h13"/><path d="M8 12h13"/><path d="M8 18h13"/><path d="M3 6h.01"/><path d="M3 12h.01"/><path d="M3 18h.01"/></g></g>
<text x="38" y="86" font-size="14" class="dg-text" fill="#17171c" font-weight="700">Preparation</text>
<text x="38" y="106" font-size="12" class="dg-muted" fill="#5c5c68" font-weight="500">Roles, contacts, playbooks,</text>
<text x="38" y="123" font-size="12" class="dg-muted" fill="#5c5c68" font-weight="500">logging, backups, training</text>
<path class="dg-line" d="M252.5 89H274.5" stroke="#cfcfc8" stroke-width="2.5" fill="none" stroke-linecap="round" stroke-linejoin="round"/>
<path class="dg-line" d="M266.5 83l8 6-8 6" stroke="#cfcfc8" stroke-width="2.5" fill="none" stroke-linecap="round" stroke-linejoin="round"/>
<rect class="dg-card" x="280.5" y="24" width="222.5" height="130" rx="12" fill="#ffffff" stroke="#cfcfc8"/>
<g class="dg-l1"><circle class="dg-solid" cx="302.5" cy="48" r="12" fill="#0369a1"/><text x="302.5" y="52.5" font-size="12.5" class="dg-on" fill="#ffffff" font-weight="700" text-anchor="middle">2</text></g>
<g class="dg-l1"><g class="dg-icon" transform="translate(463 36) scale(1.0833)" fill="none" stroke="#0369a1" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round"><path d="M4 6h16"/><path d="M4 12h10"/><path d="M4 18h6"/><circle cx="18" cy="16" r="3"/><path d="m22 20-1.8-1.8"/></g></g>
<text x="294.5" y="86" font-size="14" class="dg-text" fill="#17171c" font-weight="700">Detection and analysis</text>
<text x="294.5" y="106" font-size="12" class="dg-muted" fill="#5c5c68" font-weight="500">Spot events, confirm an</text>
<text x="294.5" y="123" font-size="12" class="dg-muted" fill="#5c5c68" font-weight="500">incident, triage, prioritise</text>
<path class="dg-line" d="M509 89H531" stroke="#cfcfc8" stroke-width="2.5" fill="none" stroke-linecap="round" stroke-linejoin="round"/>
<path class="dg-line" d="M523 83l8 6-8 6" stroke="#cfcfc8" stroke-width="2.5" fill="none" stroke-linecap="round" stroke-linejoin="round"/>
<rect class="dg-card" x="537" y="24" width="222.5" height="130" rx="12" fill="#ffffff" stroke="#cfcfc8"/>
<g class="dg-l1"><circle class="dg-solid" cx="559" cy="48" r="12" fill="#0369a1"/><text x="559" y="52.5" font-size="12.5" class="dg-on" fill="#ffffff" font-weight="700" text-anchor="middle">3</text></g>
<g class="dg-l1"><g class="dg-icon" transform="translate(719.5 36) scale(1.0833)" fill="none" stroke="#0369a1" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round"><path d="M20 13c0 5-3.5 7.5-7.66 8.95a1 1 0 0 1-.67-.01C7.5 20.5 4 18 4 13V6a1 1 0 0 1 1-1c2 0 4.5-1.2 6.24-2.72a1.17 1.17 0 0 1 1.52 0C14.51 3.81 17 5 19 5a1 1 0 0 1 1 1z"/></g></g>
<text x="551" y="86" font-size="14" class="dg-text" fill="#17171c" font-weight="700">Containment</text>
<text x="551" y="106" font-size="12" class="dg-muted" fill="#5c5c68" font-weight="500">Eradication and recovery: stop</text>
<text x="551" y="123" font-size="12" class="dg-muted" fill="#5c5c68" font-weight="500">the spread, remove the cause</text>
<path class="dg-line" d="M765.5 89H787.5" stroke="#cfcfc8" stroke-width="2.5" fill="none" stroke-linecap="round" stroke-linejoin="round"/>
<path class="dg-line" d="M779.5 83l8 6-8 6" stroke="#cfcfc8" stroke-width="2.5" fill="none" stroke-linecap="round" stroke-linejoin="round"/>
<rect class="dg-card" x="793.5" y="24" width="222.5" height="130" rx="12" fill="#ffffff" stroke="#cfcfc8"/>
<g class="dg-l1"><circle class="dg-solid" cx="815.5" cy="48" r="12" fill="#0369a1"/><text x="815.5" y="52.5" font-size="12.5" class="dg-on" fill="#ffffff" font-weight="700" text-anchor="middle">4</text></g>
<g class="dg-l1"><g class="dg-icon" transform="translate(976 36) scale(1.0833)" fill="none" stroke="#0369a1" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round"><path d="M15 2H6a2 2 0 0 0-2 2v16a2 2 0 0 0 2 2h12a2 2 0 0 0 2-2V7Z"/><path d="M14 2v4a2 2 0 0 0 2 2h4"/><path d="M10 9H8"/><path d="M16 13H8"/><path d="M16 17H8"/></g></g>
<text x="807.5" y="86" font-size="14" class="dg-text" fill="#17171c" font-weight="700">Post-incident activity</text>
<text x="807.5" y="106" font-size="12" class="dg-muted" fill="#5c5c68" font-weight="500">Lessons learned, report,</text>
<text x="807.5" y="123" font-size="12" class="dg-muted" fill="#5c5c68" font-weight="500">evidence retention</text>
<g class="dg-l2"><path class="dg-stroke" d="M904.75 156V194H135.25V160" stroke="#15803d" stroke-width="2" fill="none" stroke-linecap="round" stroke-linejoin="round"/></g>
<g class="dg-l2"><path class="dg-stroke" d="M129.25 171l6-8 6 8" stroke="#15803d" stroke-width="2.5" fill="none" stroke-linecap="round" stroke-linejoin="round"/></g>
<g class="dg-l2"><rect class="dg-bg-plain" x="329.56" y="182" width="380.88" height="24" rx="12" fill="#ffffff"/><text x="520" y="198.5" font-size="12" class="dg-ink" fill="#15803d" font-weight="600" text-anchor="middle">Lessons learned improve preparation: the cycle repeats</text></g>
<text x="520" y="238" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500" text-anchor="middle">NIST SP 800-61r2 (2012), now superseded; still a useful mental model</text>
</svg></figure>

*Figure 1. The four phases of SP 800-61r2. Real incidents loop back: analysis often continues during containment.*

**Revision 3 (April 2025)** supersedes revision 2. It is titled *Incident Response Recommendations and Considerations for Cybersecurity Risk Management: A CSF 2.0 Community Profile*. The big change is the frame. Instead of a stand-alone four-phase cycle, it organises incident response around the six functions of the **NIST Cybersecurity Framework (CSF) 2.0**, released in February 2024:

| CSF 2.0 function | What it means for incident response |
|---|---|
| **Govern** | Strategy, policy, roles and risk appetite: who decides, what is acceptable, who is accountable |
| **Identify** | Know your assets and risks, and find improvements, including those from past incidents |
| **Protect** | Safeguards that reduce the chance and the impact: access control, logging, backups, training |
| **Detect** | Find and analyse possible attacks and compromises |
| **Respond** | Take action on a detected incident: triage, contain, eradicate, communicate, report |
| **Recover** | Restore assets and normal operations |

The idea behind it: incident response is part of managing cybersecurity risk, not a separate team activity. Govern, Identify and Protect are the *preparation* side. Detect, Respond and Recover are what you do *during* an incident. Lessons learned flow back into Identify and Govern as improvements. The functions are not strict sequential steps: you may be detecting new activity while you respond, and recovery planning starts early.

<figure class="diagram" style="--dg-w: 1120px"><svg xmlns="http://www.w3.org/2000/svg" class="dg" viewBox="0 0 1120 408" width="1120" height="408" role="img" aria-labelledby="ir-csf-functions-title ir-csf-functions-desc" font-family="Inter, system-ui, -apple-system, Segoe UI, Roboto, sans-serif" xml:lang="en">
<title id="ir-csf-functions-title">The CSF 2.0 functions and incident response in NIST SP 800-61r3</title>
<desc id="ir-csf-functions-desc">A wide bar at the top for the Govern function: strategy, roles, policy and risk management, which frame everything below. Underneath, five function cards in two groups. The group Getting ready, which corresponds to preparation in the older r2 model, holds Identify (assets, risks and improvement areas) and Protect (safeguards, access, backups, training). The group Handling an incident, which corresponds to detection through recovery in r2, holds Detect (find and analyse possible incidents), Respond (contain, eradicate, communicate, report) and Recover (restore services and normal operation). A return arrow from the second group to the first is labelled: lessons learned feed back into Identify and Govern.</desc>
<rect class="dg-bg" x="0.5" y="0.5" width="1119" height="407" rx="16" fill="#ffffff" stroke="#e3e3dd"/>
<g class="dg-l7"><rect class="dg-tint" x="24" y="24" width="1072" height="64" rx="12" fill="#f3eefc" stroke="#bd9eee" stroke-width="1.2"/><text x="94" y="52" font-size="15" class="dg-ink" fill="#6d28d9" font-weight="700">GOVERN</text><text x="94" y="72" font-size="13" class="dg-text" fill="#17171c" font-weight="500">Strategy, roles, policy and risk management: it frames everything below</text></g>
<g class="dg-l7"><g class="dg-icon" transform="translate(44 40) scale(1.3333)" fill="none" stroke="#6d28d9" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round"><path d="M6 22V4a2 2 0 0 1 2-2h8a2 2 0 0 1 2 2v18Z"/><path d="M6 12H4a2 2 0 0 0-2 2v6a2 2 0 0 0 2 2h2"/><path d="M18 9h2a2 2 0 0 1 2 2v9a2 2 0 0 1-2 2h-2"/><path d="M10 6h4"/><path d="M10 10h4"/><path d="M10 14h4"/><path d="M10 18h4"/></g></g>
<g class="dg-l1"><rect class="dg-tint" x="24" y="116" width="422.4" height="214" rx="16" fill="#f0f6f9" stroke="#9ac3d9" stroke-width="1.2" stroke-dasharray="6 5"/>
<g class="dg-icon" transform="translate(40 129) scale(0.7500)" fill="none" stroke="#0369a1" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><path d="M8 6h13"/><path d="M8 12h13"/><path d="M8 18h13"/><path d="M3 6h.01"/><path d="M3 12h.01"/><path d="M3 18h.01"/></g>
<text x="66" y="143" font-size="13" class="dg-ink" fill="#0369a1" font-weight="700">Getting ready (r2: preparation)</text></g>
<g class="dg-l4"><rect class="dg-tint" x="470.4" y="116" width="625.6" height="214" rx="16" fill="#fbf4f0" stroke="#e7b39e" stroke-width="1.2" stroke-dasharray="6 5"/>
<g class="dg-icon" transform="translate(486.4 129) scale(0.7500)" fill="none" stroke="#c2410c" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><path d="m21.73 18-8-14a2 2 0 0 0-3.48 0l-8 14A2 2 0 0 0 4 21h16a2 2 0 0 0 1.73-3"/><path d="M12 9v4"/><path d="M12 17h.01"/></g>
<text x="512.4" y="143" font-size="13" class="dg-ink" fill="#c2410c" font-weight="700">Handling an incident (r2: detection to recovery)</text></g>
<rect class="dg-card" x="40" y="166" width="187.2" height="138" rx="12" fill="#ffffff" stroke="#cfcfc8"/>
<g class="dg-l1"><g class="dg-icon" transform="translate(56 182) scale(1.2500)" fill="none" stroke="#0369a1" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round"><path d="M8 6h13"/><path d="M8 12h13"/><path d="M8 18h13"/><path d="M3 6h.01"/><path d="M3 12h.01"/><path d="M3 18h.01"/></g></g>
<g class="dg-l1"><text x="98" y="203" font-size="15" class="dg-ink" fill="#0369a1" font-weight="700">Identify</text></g>
<text x="58" y="234" font-size="13" class="dg-muted" fill="#5c5c68" font-weight="500">Assets, risks and</text>
<text x="58" y="254" font-size="13" class="dg-muted" fill="#5c5c68" font-weight="500">improvement areas</text>
<rect class="dg-card" x="243.2" y="166" width="187.2" height="138" rx="12" fill="#ffffff" stroke="#cfcfc8"/>
<g class="dg-l1"><g class="dg-icon" transform="translate(259.2 182) scale(1.2500)" fill="none" stroke="#0369a1" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round"><rect x="4" y="11" width="16" height="10" rx="2"/><path d="M8 11V7a4 4 0 0 1 8 0v4"/></g></g>
<g class="dg-l1"><text x="301.2" y="203" font-size="15" class="dg-ink" fill="#0369a1" font-weight="700">Protect</text></g>
<text x="261.2" y="234" font-size="13" class="dg-muted" fill="#5c5c68" font-weight="500">Safeguards, access,</text>
<text x="261.2" y="254" font-size="13" class="dg-muted" fill="#5c5c68" font-weight="500">backups, training</text>
<rect class="dg-card" x="486.4" y="166" width="187.2" height="138" rx="12" fill="#ffffff" stroke="#cfcfc8"/>
<g class="dg-l4"><g class="dg-icon" transform="translate(502.4 182) scale(1.2500)" fill="none" stroke="#c2410c" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round"><path d="M4 6h16"/><path d="M4 12h10"/><path d="M4 18h6"/><circle cx="18" cy="16" r="3"/><path d="m22 20-1.8-1.8"/></g></g>
<g class="dg-l4"><text x="544.4" y="203" font-size="15" class="dg-ink" fill="#c2410c" font-weight="700">Detect</text></g>
<text x="504.4" y="234" font-size="13" class="dg-muted" fill="#5c5c68" font-weight="500">Find and analyse</text>
<text x="504.4" y="254" font-size="13" class="dg-muted" fill="#5c5c68" font-weight="500">possible incidents</text>
<rect class="dg-card" x="689.6" y="166" width="187.2" height="138" rx="12" fill="#ffffff" stroke="#cfcfc8"/>
<g class="dg-l4"><g class="dg-icon" transform="translate(705.6 182) scale(1.2500)" fill="none" stroke="#c2410c" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round"><path d="M20 13c0 5-3.5 7.5-7.66 8.95a1 1 0 0 1-.67-.01C7.5 20.5 4 18 4 13V6a1 1 0 0 1 1-1c2 0 4.5-1.2 6.24-2.72a1.17 1.17 0 0 1 1.52 0C14.51 3.81 17 5 19 5a1 1 0 0 1 1 1z"/></g></g>
<g class="dg-l4"><text x="747.6" y="203" font-size="15" class="dg-ink" fill="#c2410c" font-weight="700">Respond</text></g>
<text x="707.6" y="234" font-size="13" class="dg-muted" fill="#5c5c68" font-weight="500">Contain, eradicate,</text>
<text x="707.6" y="254" font-size="13" class="dg-muted" fill="#5c5c68" font-weight="500">communicate, report</text>
<rect class="dg-card" x="892.8" y="166" width="187.2" height="138" rx="12" fill="#ffffff" stroke="#cfcfc8"/>
<g class="dg-l4"><g class="dg-icon" transform="translate(908.8 182) scale(1.2500)" fill="none" stroke="#c2410c" stroke-width="1.8" stroke-linecap="round" stroke-linejoin="round"><path d="M20 6 9 17l-5-5"/></g></g>
<g class="dg-l4"><text x="950.8" y="203" font-size="15" class="dg-ink" fill="#c2410c" font-weight="700">Recover</text></g>
<text x="910.8" y="234" font-size="13" class="dg-muted" fill="#5c5c68" font-weight="500">Restore services and</text>
<text x="910.8" y="254" font-size="13" class="dg-muted" fill="#5c5c68" font-weight="500">normal operation</text>
<g class="dg-l2"><path class="dg-stroke" d="M783.2 332V364H235.2V336" stroke="#15803d" stroke-width="2" fill="none" stroke-linecap="round" stroke-linejoin="round"/></g>
<g class="dg-l2"><path class="dg-stroke" d="M229.2 347l6-8 6 8" stroke="#15803d" stroke-width="2.5" fill="none" stroke-linecap="round" stroke-linejoin="round"/></g>
<g class="dg-l2"><text x="509.2" y="388" font-size="12.5" class="dg-ink" fill="#15803d" font-weight="600" text-anchor="middle">Lessons learned feed back into Identify and Govern (r2: post-incident activity)</text></g>
</svg></figure>

*Figure 2. How the older four phases map onto the CSF 2.0 functions. The mapping is approximate: it is a way to relate the two models, not an official table.*

So which should you learn? Both. The r2 phases remain the clearest way to describe *what you do hour by hour*, and most playbooks, tools and exams still use them. The r3 frame tells you how incident response fits into the organisation's risk management and governance. The rest of this lesson follows the r2 flow and points out where the CSF 2.0 functions apply.

> [!NOTE]
> Two caveats. First, this lesson shows the high-level relationship between the r2 phases and the CSF 2.0 functions. SP 800-61r3 maps its recommendations to individual CSF subcategories; read the publication itself when you need that level of detail. Second, the triage categories used later in this lesson (functional impact, information impact, recoverability) come from r2. Check which scale your organisation's procedures use.

> [!CHECK]
> What is the main change in NIST SP 800-61r3 compared with r2?
>
> - [ ] It removes the need for a preparation stage
> - [ ] It replaces incident response with automated tools
> - [x] It organises incident response around the CSF 2.0 functions (Govern, Identify, Protect, Detect, Respond, Recover) and treats it as part of cybersecurity risk management
> - [ ] It adds a fifth phase called "Escalation" to the four-phase cycle
>
> Preparation is still essential, and it now maps to Govern, Identify and Protect. The four-phase cycle remains a handy mental model, but r3 frames the work with the CSF 2.0 functions.

### Preparation and detection
Preparation is the only phase you can do calmly, so it decides how the rest goes. It covers:

- **Roles.** Name an incident manager (decides and coordinates), technical responders, a communications lead, a legal or privacy contact and an executive sponsor. A **computer security incident response team (CSIRT)** is a group with this job; in a small company it is a few people with hats. Write down who can authorise the drastic steps, such as taking a system offline.
- **Contacts.** A list with phone numbers, kept *outside* the systems that might fail: on paper or in a separate tool. Include internal owners, your hosting provider, your insurer, legal counsel and the authorities you may need to notify.
- **Playbooks.** A **playbook** is a short, tested checklist for one kind of incident (phishing, ransomware, lost laptop, compromised account). It says what to check first, what to decide and who to call.
- **Logging.** You can't investigate what you didn't record. Keep authentication, web server, firewall, DNS and cloud audit logs; send copies to a separate log server so an attacker who owns one machine can't erase the evidence; keep clocks in sync with the Network Time Protocol (NTP) so timestamps from different systems line up; and decide how long you keep logs.
- **Backups.** Regular, tested, and with at least one copy that an attacker on your network can't delete or encrypt (offline or immutable).
- **Practice.** A **tabletop exercise**, where the team talks through a scenario around a table, finds the gaps in a few hours.

Then **detection**. Incidents are noticed through many channels:

- Alerts from a **security information and event management (SIEM)** system, an endpoint detection and response (EDR) tool or an intrusion detection system (IDS).
- Logs that someone reads or searches, as you will do in the lab.
- Reports from users ("I got a strange email", "my files have a new extension").
- Notices from outside: a cloud provider, a customer, a researcher, a law-enforcement agency, or an attacker's ransom note.

r2 distinguishes two kinds of signal. A **precursor** is a sign that an incident may happen in future (a vulnerability scan of your network, a new exploit announced for software you run). An **indicator** is a sign that an incident may have happened or is happening (an antivirus alert, many failed logins, an unusual outbound transfer). An **indicator of compromise (IoC)** is a concrete artefact that points to a compromise: an attacker's IP address, a file hash, a malicious domain.

> [!CHECK]
> Why should a copy of the logs be sent to a separate log server?
>
> - [ ] To make searching faster on the original server
> - [ ] To save disk space on the original server
> - [x] Because an attacker who compromises a machine can edit or delete its local logs, but not the remote copy
> - [ ] Because logs are only legally valid on a dedicated server
>
> Attackers often clear local logs to hide. A remote copy that the compromised machine can't change keeps the evidence intact.

### Triage and prioritisation
You will rarely have enough people to treat every alert as a crisis. **Triage** is the quick sorting that decides what gets attention first, like hospital triage. It has two steps: validate (is this real?) and prioritise (how bad, how urgent?).

Validate by looking for a second source. A failed-login alert plus a successful login from the same address is much stronger than either alone. Document what you see, with the time and where you saw it, from the very first minute.

To prioritise, SP 800-61r2 suggests three questions. These categories come from r2; your own policy may use different labels or scales, so use them as a model.

| Factor | Question | r2 categories |
|---|---|---|
| **Functional impact** | How much does it disrupt what the organisation does? | None, Low, Medium, High |
| **Information impact** | What happened to the data? | None, Privacy breach, Proprietary breach, Integrity loss |
| **Recoverability** | How long and how much effort to get back to normal? | Regular, Supplemented, Extended, Not recoverable |

The last category is the tricky one. *Regular* means recovery time is predictable with the resources you have. *Supplemented* means predictable but you need extra resources. *Extended* means you can't predict the time and need outside help. *Not recoverable* applies, for example, to data that was stolen and published: you can restore your systems, but you can't take the data back.

The factors can disagree, and that is the point. A defaced marketing page has high visibility, low information impact and quick recovery. A quiet download of customer records has **no** functional impact (the shop works fine) but a serious privacy breach and an unrecoverable loss of confidentiality. Functional impact alone would rank it last; it should rank first.

Priority then drives the response: who is woken up, how fast, and who must be told. Write the classification down, and revisit it as you learn more.

> [!CHECK]
> An attacker silently downloads a customer database, while the website works normally. Which statement fits best?
>
> - [ ] Low priority: functional impact is none
> - [ ] Low priority: the system is easy to restore
> - [x] High priority: functional impact is none, but there is a privacy breach and the stolen data can't be recovered
> - [ ] Not an incident, because nothing was deleted
>
> Prioritisation weighs all three factors. A confidentiality loss can be the most serious kind of incident even when nothing visibly breaks.

### Evidence and the order of volatility
Everything you do during an incident changes the system. Rebooting erases memory. Logging in alters file timestamps. Yet the evidence you collect may be needed to understand the incident, to stop it happening again, to claim insurance or to support a legal case. So you collect carefully.

RFC 3227 (*Guidelines for Evidence Collection and Archiving*, Request for Comments, 2002) gives the core rules:

- Capture as accurate a picture of the system as you can, and keep detailed notes with dates and times, including any difference between the system clock and the real time.
- Change as little as possible while collecting.
- Collect first, analyse later. Work on copies.
- Proceed from the **most volatile** data to the least.
- Don't shut the system down before you've collected what you need, and prefer trusted tools (for example, from read-only media) over programs installed on a machine you don't trust.

The **order of volatility** ranks evidence by how quickly it disappears. RFC 3227 lists, from most to least volatile: registers and cache; the routing table, ARP cache, process table, kernel statistics and memory; temporary file systems; disk; remote logging and monitoring data relevant to the system; physical configuration and network topology; and archival media.

<figure class="diagram" style="--dg-w: 1040px"><svg xmlns="http://www.w3.org/2000/svg" class="dg" viewBox="0 0 1040 530" width="1040" height="530" role="img" aria-labelledby="ir-order-of-volatility-title ir-order-of-volatility-desc" font-family="Inter, system-ui, -apple-system, Segoe UI, Roboto, sans-serif" xml:lang="en">
<title id="ir-order-of-volatility-title">Order of volatility from RFC 3227</title>
<desc id="ir-order-of-volatility-desc">A table of seven rows, from most volatile at the top (collect first) to least volatile at the bottom (collect last). 1, registers and cache: last nanoseconds to milliseconds, rarely practical to capture outside lab tools. 2, routing table, ARP cache, process table, kernel statistics and memory: last seconds to minutes and are gone at power-off; capture with a memory acquisition tool and by saving command output. 3, temporary file systems: last until reboot or clean-up; copy the temporary folders. 4, disk: lasts until overwritten; take a forensic image, a bit-for-bit copy. 5, remote logging and monitoring data: lifetime set by the retention policy; export from the log server or SIEM. 6, physical configuration and network topology: months to years; diagrams, photos and notes. 7, archival media: years; secure the backups and keep them intact.</desc>
<rect class="dg-bg" x="0.5" y="0.5" width="1039" height="529" rx="16" fill="#ffffff" stroke="#e3e3dd"/>
<text x="84" y="38" font-size="11" class="dg-muted" fill="#5c5c68" font-weight="700" letter-spacing="1.2">EVIDENCE SOURCE</text>
<text x="468" y="38" font-size="11" class="dg-muted" fill="#5c5c68" font-weight="700" letter-spacing="1.2">HOW LONG IT LASTS</text>
<text x="716" y="38" font-size="11" class="dg-muted" fill="#5c5c68" font-weight="700" letter-spacing="1.2">HOW TO CAPTURE IT</text>
<g class="dg-l4"><rect class="dg-tint" x="24" y="60" width="992" height="52" rx="10" fill="#faf0ec" stroke="#e7b39e" stroke-width="1.2"/><circle class="dg-solid" cx="48" cy="86" r="12" fill="#c2410c"/><text x="48" y="90.5" font-size="12.5" class="dg-on" fill="#ffffff" font-weight="700" text-anchor="middle">1</text></g>
<text x="84" y="90.5" font-size="13" class="dg-text" fill="#17171c" font-weight="700">Registers and cache</text>
<text x="468" y="90.5" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500">nanoseconds to milliseconds</text>
<text x="716" y="90.5" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500">rarely practical outside lab tools</text>
<g class="dg-l4"><rect class="dg-tint" x="24" y="120" width="992" height="52" rx="10" fill="#faf0ec" stroke="#e7b39e" stroke-width="1.2"/><circle class="dg-solid" cx="48" cy="146" r="12" fill="#c2410c"/><text x="48" y="150.5" font-size="12.5" class="dg-on" fill="#ffffff" font-weight="700" text-anchor="middle">2</text></g>
<text x="84" y="142" font-size="13" class="dg-text" fill="#17171c" font-weight="700">Routing table, ARP cache, process table,</text>
<text x="84" y="160" font-size="13" class="dg-text" fill="#17171c" font-weight="700">kernel statistics, memory</text>
<text x="468" y="150.5" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500">seconds to minutes; gone at power-off</text>
<text x="716" y="150.5" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500">memory acquisition tool; save command output</text>
<g class="dg-l3"><rect class="dg-tint" x="24" y="180" width="992" height="52" rx="10" fill="#f9f1eb" stroke="#e1ba9d" stroke-width="1.2"/><circle class="dg-solid" cx="48" cy="206" r="12" fill="#b45309"/><text x="48" y="210.5" font-size="12.5" class="dg-on" fill="#ffffff" font-weight="700" text-anchor="middle">3</text></g>
<text x="84" y="210.5" font-size="13" class="dg-text" fill="#17171c" font-weight="700">Temporary file systems</text>
<text x="468" y="210.5" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500">until reboot or clean-up</text>
<text x="716" y="210.5" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500">copy the temporary folders</text>
<g class="dg-l3"><rect class="dg-tint" x="24" y="240" width="992" height="52" rx="10" fill="#f9f1eb" stroke="#e1ba9d" stroke-width="1.2"/><circle class="dg-solid" cx="48" cy="266" r="12" fill="#b45309"/><text x="48" y="270.5" font-size="12.5" class="dg-on" fill="#ffffff" font-weight="700" text-anchor="middle">4</text></g>
<text x="84" y="270.5" font-size="13" class="dg-text" fill="#17171c" font-weight="700">Disk</text>
<text x="468" y="270.5" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500">until overwritten</text>
<text x="716" y="270.5" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500">forensic image (bit-for-bit copy)</text>
<g class="dg-l2"><rect class="dg-tint" x="24" y="300" width="992" height="52" rx="10" fill="#ecf5ef" stroke="#a1ccb1" stroke-width="1.2"/><circle class="dg-solid" cx="48" cy="326" r="12" fill="#15803d"/><text x="48" y="330.5" font-size="12.5" class="dg-on" fill="#ffffff" font-weight="700" text-anchor="middle">5</text></g>
<text x="84" y="330.5" font-size="13" class="dg-text" fill="#17171c" font-weight="700">Remote logging and monitoring data</text>
<text x="468" y="330.5" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500">set by the retention policy</text>
<text x="716" y="330.5" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500">export from the log server or SIEM</text>
<g class="dg-l1"><rect class="dg-tint" x="24" y="360" width="992" height="52" rx="10" fill="#ebf3f7" stroke="#9ac3d9" stroke-width="1.2"/><circle class="dg-solid" cx="48" cy="386" r="12" fill="#0369a1"/><text x="48" y="390.5" font-size="12.5" class="dg-on" fill="#ffffff" font-weight="700" text-anchor="middle">6</text></g>
<text x="84" y="382" font-size="13" class="dg-text" fill="#17171c" font-weight="700">Physical configuration,</text>
<text x="84" y="400" font-size="13" class="dg-text" fill="#17171c" font-weight="700">network topology</text>
<text x="468" y="390.5" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500">months to years</text>
<text x="716" y="390.5" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500">diagrams, photos, notes</text>
<g class="dg-l1"><rect class="dg-tint" x="24" y="420" width="992" height="52" rx="10" fill="#ebf3f7" stroke="#9ac3d9" stroke-width="1.2"/><circle class="dg-solid" cx="48" cy="446" r="12" fill="#0369a1"/><text x="48" y="450.5" font-size="12.5" class="dg-on" fill="#ffffff" font-weight="700" text-anchor="middle">7</text></g>
<text x="84" y="450.5" font-size="13" class="dg-text" fill="#17171c" font-weight="700">Archival media</text>
<text x="468" y="450.5" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500">years</text>
<text x="716" y="450.5" font-size="12.5" class="dg-muted" fill="#5c5c68" font-weight="500">secure the backups; keep them intact</text>
<text x="24" y="504" font-size="12.5" class="dg-ink" fill="#c2410c" font-weight="700">▲ Most volatile: collect first</text>
<text x="1016" y="504" font-size="12.5" class="dg-ink" fill="#0369a1" font-weight="700" text-anchor="end">▼ Least volatile: collect last</text>
</svg></figure>

*Figure 3. The most fragile evidence first: a running system tells you things that a powered-off one never will.*

A **forensic image** is a bit-for-bit copy of a disk or memory. After copying, compute a cryptographic hash (SHA-256 is a good choice today) and write it down. If the hash of the copy still matches later, nobody has altered it.

**Chain of custody** is the written record of who handled the evidence. For each item, record where, when and by whom it was found and collected; who has had custody and when; where it is stored and how it is protected; and every transfer, with date, time and the people involved. A gap in the chain gives anyone a reason to doubt the evidence. You can do it with a simple table or form, so long as you do it every time.

> [!CHECK]
> A suspect server is still running. Which evidence should you collect first?
>
> - [ ] The disk image, because it holds the most data
> - [ ] The archived backups, because they are the safest
> - [x] The running connections, processes and memory, because they vanish when the system is powered off
> - [ ] Nothing: shut the server down at once to stop the attacker
>
> The order of volatility says to capture what disappears fastest first. Powering off loses memory and network state, and should come after collection (or after a deliberate decision to accept that loss).

### Containment, eradication and recovery
Now you act. These are three separate jobs, and mixing them up is a common source of mistakes.

**Containment** stops the incident from spreading or getting worse. **Eradication** removes the cause. **Recovery** restores normal operation safely.

Containment comes in two time-scales:

- **Short-term containment** is the fast, often crude action that buys time: disable the account, block the IP address, isolate the machine from the network, cut a VPN tunnel.
- **Long-term containment** keeps the business running while you prepare a clean fix: move the service to a temporary clean server, put extra filtering in front of it, add monitoring.

The big decision is **isolate or monitor**. Isolating stops the damage now but tells the attacker you've noticed, so they may delete their tools, change tactics or retaliate. Monitoring lets you learn what the attacker does and how far they got, but every minute allows more damage. r2 lists criteria for the choice: potential damage and theft of resources; the need to preserve evidence; the service's availability needs; the time and resources the strategy needs; how effective it is; and how long the solution will last. When data is leaving, you isolate. When you're unsure how far the attacker has got, you may watch briefly with a plan to cut access.

Whatever you choose, preserve evidence *first or at the same time*: save the logs, record the running state, then change things.

**Eradication** removes whatever let the attacker in and whatever they left behind: delete malware, close the vulnerability, remove accounts or keys they created, reset every credential they could have seen. If you can't be sure a machine is clean, rebuild it from a known-good image rather than "cleaning" it.

**Recovery** returns systems to service: restore from a clean backup or rebuild, apply patches, test, bring services back in stages, and watch closely for signs that the attacker returns. Declare the incident closed only when you are confident the cause is gone.

| Action | Type | Notes |
|---|---|---|
| Disable the compromised account and kill its sessions | Containment (short-term) | Fast; also reset the password |
| Block the attacker's IP at the firewall | Containment (short-term) | Easy to bypass; do it, but don't rely on it |
| Move the app to a clean server behind extra filtering | Containment (long-term) | Keeps the business running during the fix |
| Remove an account the attacker created | Eradication | Check for other persistence too: keys, scheduled jobs |
| Rebuild the server from a trusted image | Eradication and recovery | Safer than cleaning by hand |
| Restore data, test, return to service in stages, monitor | Recovery | Closely watch the first days |

> [!CHECK]
> Data is being copied out of a server right now. What should the incident manager decide first?
>
> - [ ] Wait and monitor for a week to identify the whole attacker group
> - [x] Preserve what you can quickly, then isolate the affected account or system to stop the loss
> - [ ] Reinstall the operating system immediately, before anyone looks at it
> - [ ] Publish a statement before doing anything technical
>
> With active data loss, the cost of waiting is high, so you contain now. Taking notes and snapshots as you go preserves evidence; a reinstall would destroy it.

### Communication, notification and learning
During an incident you manage people as much as machines.

- **Keep a log of decisions and actions** with timestamps. It is your timeline and, later, the heart of the report.
- **Share on a need-to-know basis.** The fewer people who talk, the fewer wrong messages. Name one spokesperson. Don't speculate: say what you know, what you don't and when you'll update.
- **Use a separate channel** if the attacker may be reading your email or chat.
- **Know your external duties.** Contracts, insurers and laws can require you to notify someone, sometimes quickly. In the European Union (EU), the General Data Protection Regulation (GDPR) is the best-known example: a controller that suffers a personal data breach must notify the supervisory authority without undue delay and, where feasible, within 72 hours of becoming aware of it, unless the breach is unlikely to result in a risk to people's rights and freedoms (Article 33). If the risk to individuals is high, they must be told too (Article 34). Other laws, sectors and countries have their own rules and clocks.
- **Bring in legal or privacy advice early.** Whether something is a notifiable breach is a legal judgement, made with the facts the technical team provides.

> [!NOTE]
> The GDPR example is a simplification for learning, not legal advice. The 72-hour clock starts when the organisation becomes *aware* of the breach, which is itself a judgement call. Other EU rules, such as the NIS2 Directive, add their own reporting stages for organisations in scope. Check the law that applies to you, and ask your data protection officer (DPO) or legal counsel.

After the incident comes the part many teams skip: learning. A **blameless post-incident review** is a meeting, held within days while memories are fresh, whose goal is to improve the system, not to find someone to punish. People who fear blame hide mistakes, and hidden mistakes repeat. Ask:

- What happened, minute by minute? (Use the timeline.)
- What was the **root cause**, the underlying condition that, if fixed, would have prevented it? "Dave chose a weak password" is a symptom. "Password-only SSH was open to the Internet with no lockout and no second factor" is closer to a root cause. Contributing factors usually outnumber root causes.
- What worked? What slowed us down?
- What will we change, who owns each action and by when?

Then write the **incident report**. A short one contains:

| Section | Content |
|---|---|
| Summary | What happened, in two or three sentences for a non-technical reader |
| Timeline | Key events with date, time and time zone |
| Impact | Systems, data and people affected; downtime; notifications made |
| Root cause and contributing factors | Why it was possible |
| Response | What was done: detection, containment, eradication, recovery |
| Lessons learned | What worked and what didn't |
| Actions | Fixes, with an owner and a due date for each |

> [!CHECK]
> In a blameless post-incident review, which finding is closest to a root cause?
>
> - [ ] "An employee reused a weak password"
> - [ ] "The on-call engineer replied too slowly at 3 a.m."
> - [x] "The admin pages accepted a password alone from any Internet address, with no lockout or second factor"
> - [ ] "The attacker was skilled"
>
> A root cause is a condition in the system that you can fix. The password choice and the skill of the attacker are real, but fixing them is not within the organisation's control; the missing controls are.

## How it works
Putting the pieces together, a typical incident runs like this. The steps overlap in real life, and you will often go back a step.

1. **Detect.** A signal arrives: an alert, a log search, a user report.
2. **Validate and record.** Is this real? Start the notes: what, when, who saw it. Look for a second source.
3. **Triage.** Classify by functional impact, information impact and recoverability. Assign an incident manager.
4. **Preserve.** Copy the logs, record volatile state, hash it, start the chain of custody.
5. **Analyse and scope.** Build the timeline. Which accounts, hosts and data? How did they get in? Are they still in?
6. **Contain.** Short-term first; long-term if you need a bridge to the fix.
7. **Eradicate.** Remove access, persistence and the weakness.
8. **Recover.** Restore, test, return gradually, monitor.
9. **Communicate.** In parallel with all of the above: management, legal, affected people, authorities.
10. **Review and improve.** Report, blameless review, action list, and update playbooks and controls.

In CSF 2.0 terms: steps 1 to 2 are Detect, 3 to 9 are Respond and Recover, and step 10 feeds Identify and Govern.

## Worked example
Here is a three-line excerpt from an invented web server `web01.example.com`, with times in Coordinated Universal Time (UTC).

```text
Nov  4 02:11:05 web01 sshd[2186]: Failed password for invalid user admin from 203.0.113.45 port 47886 ssh2
Nov  4 02:14:28 web01 sshd[2926]: Accepted password for dave from 203.0.113.45 port 50280 ssh2
203.0.113.45 - dave [04/Nov/2026:02:20:33 +0000] "GET /admin/export/customers.csv HTTP/1.1" 200 4812034 "-" "curl/7.88.1"
```

Read it as an analyst:

- **Line 1:** a failed login for a user that doesn't exist (`admin`) from `203.0.113.45`. By itself, an event.
- **Line 2:** three minutes later, the same address logs in successfully as `dave`. In the full log, that address had 23 failures first. Trying a few passwords against many accounts is called **password spraying**. Now it is an incident: probably a compromised account.
- **Line 3:** the same address, with `dave` as the authenticated user, downloads `customers.csv` (4,812,034 bytes, status 200 means it succeeded). `curl` as the client tool, rather than a browser, suggests a script.

Classification: functional impact **None** (the site works); information impact **Privacy breach** (customer data, to be confirmed by checking the file); recoverability **Not recoverable** for the stolen data. Priority: **high**, and the privacy officer should be told at once, because a notification clock may have started. You will work this scenario, with full logs, in the lab.

## Common pitfalls
- **Rebooting or "cleaning up" first.** Wiping the machine feels responsible but destroys the evidence you need to learn how the attacker got in, so they come back by the same door. Preserve, then act.
- **Fixating on the first IP address.** Blocking `203.0.113.45` is easy and feels like a fix. Attackers change address in seconds. The account, the weakness and the data matter more.
- **Ranking by downtime only.** Silent data theft has no functional impact. Always weigh information impact and recoverability too.
- **Trusting the system's own tools.** On a compromised host, commands and logs may have been altered. Compare with logs held elsewhere and use trusted tools.
- **Mixed time zones.** One log in UTC and another in local time breaks your timeline. Note each source's zone and offset.
- **No notes.** If it isn't written down with a time, it didn't happen, as far as the report and the lawyers are concerned.
- **Declaring victory too early.** If you never found how they got in, you haven't eradicated anything.
- **Blame.** A review that punishes the person who clicked teaches everyone to stay quiet next time.

## Security considerations
Incident response is itself a target and a risk.

- **Protect the logs and the evidence.** Send logs off-host, restrict who can read and delete them, and hash what you collect. Logs often contain personal data, so apply least privilege and retention limits.
- **Assume the attacker may be watching.** If an administrator's account or mailbox is compromised, the attacker can read your response plan. Use an out-of-band channel and a pre-agreed way to confirm identity.
- **Insider incidents need special care.** If the suspect may be a colleague, involve human resources and legal early and keep the circle small.
- **Stay authorised.** Only investigate, scan or collect from systems you own or are authorised to examine. "Hacking back" at an attacker's systems is illegal in most places and can harm innocent third parties, because the attacker's machines are often someone else's compromised computers.
- **Handle personal data lawfully.** Collecting evidence from employees' devices or mailboxes can have privacy and labour-law constraints. Get advice before you do it.
- **Defend in depth so the next incident is smaller.** Multi-factor authentication (MFA), rate limiting and lockout, least privilege and tested backups limit both the chance and the impact.

## Summary
- An event is any observable occurrence; an incident is an event that actually or imminently harms confidentiality, integrity or availability, or breaks policy or law.
- NIST SP 800-61r2 gave a four-phase life cycle; SP 800-61r3 (April 2025) supersedes it and organises incident response around the CSF 2.0 functions Govern, Identify, Protect, Detect, Respond and Recover. The four phases remain a good mental model.
- Preparation (roles, contacts, playbooks, logging, backups, practice) decides how well you respond.
- Triage weighs functional impact, information impact and recoverability (r2 categories). Silent data theft can be the highest priority.
- Collect evidence from the most volatile to the least (RFC 3227), work on copies, hash them and keep a chain of custody.
- Contain first (short-term, then long-term), then eradicate the cause, then recover, and decide deliberately between isolating and monitoring.
- Know your notification duties early; the GDPR's 72-hour rule for personal data breaches is one EU example.
- Close with a blameless review and a short report: summary, timeline, impact, root cause, lessons and owned actions.

## Glossary
| Term | Definition |
|---|---|
| Event | Any observable occurrence in a system or network |
| Incident | An event, or set of events, that actually or imminently harms confidentiality, integrity or availability, or violates security policy or law |
| Incident response | The organised process of preparing for, detecting, handling and learning from incidents |
| NIST CSF 2.0 | The NIST Cybersecurity Framework version 2.0 (2024), organised in six functions: Govern, Identify, Protect, Detect, Respond, Recover |
| CSIRT | Computer security incident response team: the people responsible for handling incidents |
| Playbook | A short, tested checklist for handling one type of incident |
| Tabletop exercise | A discussion-based drill in which a team talks through an incident scenario |
| SIEM | Security information and event management: a system that collects logs and raises alerts |
| Precursor | A sign that an incident may occur in the future |
| Indicator | A sign that an incident may have occurred or is occurring |
| Indicator of compromise (IoC) | A concrete artefact, such as an IP address, file hash or domain, that points to a compromise |
| Triage | Quickly validating and sorting signals to decide what to handle first |
| Functional impact | How much an incident disrupts the organisation's functions (r2: None, Low, Medium, High) |
| Information impact | What happened to the data (r2: None, Privacy breach, Proprietary breach, Integrity loss) |
| Recoverability | The time and effort needed to return to normal (r2: Regular, Supplemented, Extended, Not recoverable) |
| Order of volatility | Ranking of evidence from the quickest to disappear to the most durable (RFC 3227) |
| Forensic image | A bit-for-bit copy of a disk or memory, verified with a hash |
| Chain of custody | The written record of who collected, held and transferred each piece of evidence, and when |
| Containment | Actions that stop an incident from spreading or worsening |
| Eradication | Removing the cause of the incident and what the attacker left behind |
| Recovery | Safely restoring systems and services to normal operation |
| Blameless post-incident review | A review that focuses on improving systems and processes, not on punishing people |
| Root cause | The underlying condition that, if fixed, would have prevented the incident |
| Password spraying | Trying a few common passwords against many accounts, to avoid lockouts |

## Check your understanding
1. Explain in your own words the difference between an event and an incident. Give one event that could turn into an incident, and say what additional fact would change it.
2. Why did NIST SP 800-61r3 move to the CSF 2.0 functions? Which functions cover "getting ready" and which cover "handling an incident"?
3. Pick an incident you have read about. Classify it by functional impact, information impact and recoverability, and say which factor should drive the priority.
4. A colleague wants to wipe a suspect laptop right now "to be safe". What do you tell them, and what do you do first?
5. Write two sentences, one root cause and one symptom, for an incident where a former employee's account was still active.

## References
- [NIST SP 800-61r3](https://csrc.nist.gov/pubs/sp/800/61/r3/final): Incident Response Recommendations and Considerations for Cybersecurity Risk Management (April 2025).
- [NIST SP 800-61r2](https://csrc.nist.gov/pubs/sp/800/61/r2/final): Computer Security Incident Handling Guide (2012, superseded; source of the four phases and the triage categories).
- [NIST Cybersecurity Framework (CSF) 2.0](https://www.nist.gov/cyberframework).
- [RFC 3227](https://www.rfc-editor.org/rfc/rfc3227): Guidelines for Evidence Collection and Archiving.
- [GDPR, Articles 33 and 34](https://eur-lex.europa.eu/eli/reg/2016/679/oj): notification of a personal data breach to the supervisory authority and to the data subject.