[reference] — from the roadmap.sh cyber-security roadmap. What you do when (not if) something gets through — the disciplined process that turns a breach from a catastrophe into a contained, learned-from event.
Why a process, not panic
During a live incident, ad-hoc reactions make things worse: tipping off the attacker, destroying evidence, or taking down more than necessary. A defined incident response (IR) process — decided calmly in advance — is what lets a team act fast and correctly under pressure. It’s the security counterpart to a DevOps on-call runbook.
The IR lifecycle (NIST / SANS)
The standard phases (SANS’s PICERL is a common mnemonic):
- Preparation — before anything happens: an IR plan, a trained team with defined roles, tooling (logging/EDR in place so you have evidence), communication plans, and runbooks for common scenarios. The most important phase — you can’t improvise this mid-crisis.
- Identification / Detection — recognize that an incident is happening and determine scope: what’s affected, how bad, what kind. Triage a SIEM alert or hunt finding from “event” to “incident.”
- Containment — stop the bleeding without destroying evidence. Short-term (isolate the affected host from the network) and long-term (temporary fixes to keep operating while you eradicate). The judgment call: contain fast, but don’t tip off an attacker before you understand the full scope (they may burrow deeper).
- Eradication — remove the threat: delete malware, close the vulnerability that was exploited, reset compromised credentials, remove attacker persistence/backdoors. Half-eradication means they’re back tomorrow.
- Recovery — restore systems to normal operation safely: rebuild from known-good backups (the ransomware answer), verify systems are clean, monitor closely for the attacker’s return, and gradually return to production.
- Lessons Learned — a blameless post-incident review: what happened, how well the response went, and what to change (a new detection rule, a patch, a process fix) so it doesn’t recur. This closes the loop back into Preparation — the phase teams most often skip and most regret skipping.
Preparation → Identification → Containment → Eradication → Recovery → Lessons Learned
↑___________________________________________________________________|
(lessons feed back into preparation)
SOAR and runbooks
- Runbooks / playbooks — step-by-step procedures for specific incident types (ransomware, phishing, data breach), so responders follow a proven path instead of improvising. Preparation made concrete.
- SOAR (Security Orchestration, Automation and Response) — automates parts of the response: auto-enrich an alert with threat intel, auto-isolate a flagged host, auto-open a ticket. Fights alert volume and speeds the repetitive early steps, freeing analysts for judgment calls.
The human and legal side
IR isn’t purely technical: communication (who’s informed, when — leadership, legal, PR, customers), legal/regulatory obligations (breach-notification laws like GDPR’s 72-hour rule — compliance), and chain of custody for any evidence that might be used legally (forensics). Mishandling these can turn a well-handled technical response into a legal or reputational disaster.
Related
- Forensics & Malware Analysis — the deep investigation during/after an incident
- Logging, SIEM & Detection — where incidents are first spotted
- Governance, Risk & Compliance — the obligations IR must satisfy
Software engineering
Infrastructure & security
Computer science
Hardware & engineering