open atlas
↑ Back to track
Defensive Security BLUE · 02 · 01

The IR lifecycle

NIST splits incident response into four phases, but the one nobody is watching during an outage — Preparation — quietly decides the outcome of the other three. This is why the work you do before the pager fires is the work that saves you.

BLUE Middle ◷ 15 min
Level
FoundationsJuniorMiddleSenior

2:47 a.m. An alert fires: a service account is reading rows it has never touched, fast. The on-call engineer is awake, competent, and completely stuck. Who declares an incident? Where is the runbook? Does anyone have the cloud permissions to pull the IAM logs, and is anyone even allowed to power off a production node? They burn forty minutes paging people who don’t answer, find the wiki page is a year stale, and watch the attacker finish exfiltration while the team argues about whether this is “really” an incident. Nothing that happened at 2:47 was a technical failure. Every one of those forty minutes was lost in a decision that should have been made — and written down — months earlier, on a calm afternoon, by someone who is now asleep.

By the end of this lesson you’ll know the four NIST incident-response phases, how they feed back into each other, and why the phase that runs before any incident is the one that decides whether the other three go well or badly.

The model: four phases, not a straight line

When people picture incident response they picture the middle of it — the war room, the bridge call, the frantic typing. NIST SP 800-61 (the field’s reference, Computer Security Incident Handling Guide) frames it more usefully as a lifecycle of four phases that loop, not a script you run once top to bottom:

  1. Preparation — everything you build before an incident: the response team and its authority, runbooks, log pipelines, tooling, retained backups, contact lists, and the practice runs that make all of it real.
  2. Detection & Analysis — noticing that something happened and figuring out what: triaging an alert, confirming it’s a true incident, scoping how far it reaches, and assigning severity.
  3. Containment, Eradication & Recovery — stopping the bleeding, removing the attacker’s access and artifacts, then restoring service to a known-good state and confirming it holds.
  4. Post-Incident Activity — the “lessons learned” pass: a blameless review that turns what you just lived through into improvements.

The arrow that matters most is the one that loops the last phase back to the first. Post-Incident Activity feeds Preparation. A finding from one incident — “we had no log of service-account reads” — becomes a Preparation task that makes the next Detection phase faster. An IR program that skips the loop relives the same incident every quarter; one that closes it gets measurably harder to breach over time.

Why Preparation decides the outcome

Here is the senior insight that the four-box diagram hides: the phases are not equal in leverage. During an actual incident, the clock is running and your options are fixed — you can only respond with the access, logs, tooling, and authority that already exist. You cannot decide during the breach to start retaining the logs you’ll need to scope it. You cannot grant yourself the cloud permission to isolate a node while the attacker is using it. You cannot write, socialize, and rehearse a runbook at 2:47 a.m. Every capability you reach for under pressure was either built in Preparation or it simply isn’t there.

This is why mature teams measure two numbers: MTTD (mean time to detect) and MTTR (mean time to respond/recover). Both are dominated by Preparation. MTTD is short only if you were already logging the right events with the right context and had detections watching them — that’s the previous unit’s work, cashed in here. MTTR is short only if the runbook exists, the responder has standing authority to act, and the recovery path (clean backups, infrastructure-as-code to rebuild) was tested before the day you need it. The breach in the Hook lost forty minutes not to a hard technical problem but to missing Preparation: no clear incident commander, a stale runbook, no pre-granted permissions. The attack succeeded in the gaps the team never closed on a calm afternoon.

Why this works

Why does NIST put practice inside Preparation, not treat it as optional polish? Because an untested plan is a hypothesis, and incidents are a terrible time to test hypotheses. Teams run tabletop exercises (talk through a scenario around a table) and game days (inject a real failure and respond live) precisely to discover the gaps cheaply: the contact list is wrong, the runbook assumes a tool that was decommissioned, nobody actually has the IAM role the playbook requires. Each gap found in a drill is one you didn’t find at 2:47 a.m. with an attacker on the keyboard. A plan that has never been exercised should be assumed broken in at least one load-bearing place.

Detection, Containment, and the tradeoffs inside them

The middle phases carry their own senior judgment calls. Detection & Analysis is where you resist two opposite failures: declaring an incident on every noisy alert (alert fatigue, wasted response), or dismissing a real one as noise. The discipline is triage — confirm it’s a true positive, then scope it (what’s affected, how far, since when) before you act, because containment decisions are only as good as your understanding of blast radius. ATT&CK is the shared vocabulary that makes scoping concrete: naming the adversary’s techniques (credential access, lateral movement, exfiltration) tells you where else to look before you declare the incident bounded.

Containment hides the sharpest tradeoff in the whole lifecycle: containment versus evidence. The instinct is to pull the plug — power off the compromised box the instant you find it. But a hard power-off destroys volatile evidence (memory, live network connections, the malware’s running state) and tips the attacker off, who may have other footholds you haven’t found yet. Isolating the host on the network instead preserves forensics and may let you watch the attacker’s other moves. Yet waiting too long lets exfiltration continue. There is no universal right answer — only a judgment weighing active damage against the value of evidence and the risk of a wider, undetected presence.

Pick the best fit

You find one compromised host actively exfiltrating data, and ATT&CK-mapped signals suggest the attacker used valid credentials and may have moved laterally. Pick the best first containment move.

Quiz

Two teams have identical tooling and engineers. Team A rehearses its IR plan quarterly with tabletop exercises and game days; Team B wrote a plan two years ago and filed it. Why does Team A respond dramatically faster to a real incident?

Quiz

During a post-incident review, the team finds the breach went undetected for days because service-account database reads were never logged. In the NIST lifecycle, where does the fix belong?

Order the steps

Order the NIST SP 800-61 incident-response phases as they run, ending with the phase that loops back to the first:

  1. 1 Preparation — build the team, runbooks, logs, and tooling before anything fires
  2. 2 Detection & Analysis — confirm it's a real incident and scope the blast radius
  3. 3 Containment, Eradication & Recovery — stop it, remove access, restore to known-good
  4. 4 Post-Incident Activity — blameless review that feeds back into Preparation
Recall before you leave
  1. 01
    Name the four NIST SP 800-61 incident-response phases, what each does, and explain the feedback loop between them.
  2. 02
    Why is Preparation the highest-leverage phase, and how does it show up in MTTD and MTTR? Include the containment-versus-evidence tradeoff.
Recap

NIST SP 800-61 frames incident response as a lifecycle of four phases that loop, not a one-way script: Preparation (the team, runbooks, logs, tooling, backups, and rehearsals you build beforehand), Detection & Analysis (confirm it’s real, then scope the blast radius before acting), Containment/Eradication/Recovery (stop the bleeding, remove access, restore to known-good), and Post-Incident Activity (a blameless review). The highest-leverage edge is the feedback arrow from Post-Incident Activity back into Preparation — each incident’s findings become Preparation tasks that make the next response faster. Preparation dominates because during an incident you can only use the access, logs, authority, and tooling that already exist; both MTTD and MTTR are decided long before the pager fires. Practice is part of Preparation, not optional, because an unexercised plan is a hypothesis. And inside containment lives the sharpest tradeoff: a hard power-off stops one host but destroys evidence and tips off an attacker with other footholds, while network isolation preserves forensics and cuts egress at once. Next time you read a green incident-response plan, the senior question is: has anyone ever actually run it — and on the day it fires, will the responder already have the access and the authority to act?

Practice

Start at the top. Tasks go easiest → hardest: recall a fact, apply it to a case, then a senior-level stretch. Open one, attempt it, then reveal.

recallapplystretch0 of 6 done
Connected lessons

Something unclear?

Ask a question about this lesson. Questions are anonymous and go straight to the author to make the lesson better.

shortcuts expand
search
K
prev piece
k
next piece
j
cycle tier
t
this menu
?
sources2
expand
  1. 01
  2. 02

Trademarks belong to their respective owners. Editorial reference only.