cloud-incident-response-architect
Architect-level incident response for cloud estates — severity and declaration, roles and communication, blast-radius containment (revoke, isolate, roll back), evidence preservation, and the post-incident design changes that prevent the next one. Use when systems break in production and the response itself must be engineered.
- License 许可
- MIT
- 最近更新
- 2026-11-13
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- cloud-incident-response-architect
- license
- MIT
- metadata
- { "crusor-path": "cloud-and-devops-architect/operations-architect/cloud-incident-response-architect", "crusor-domain": "cloud-and-devops-architect", "crusor-origin": "original", "crusor-sources": "sre.google/sre-book/managing-incidents; aws.amazon.com/blogs/mt/aws-security-incident-response-guide; incident.io/blog", "crusor-triggers": "incident response, on call, severity, postmortem, outage, rollback, containment, status page, escalation, runbook", "crusor-verified-on": "2026-10-30" }
- description
- Architect-level incident response for cloud estates — severity and declaration, roles and communication, blast-radius containment (revoke, isolate, roll back), evidence preservation, and the post-incident design changes that prevent the next one. Use when systems break in production and the response itself must be engineered.
- compatibility
- Any cloud; paging, status pages, IAM revocation
03 SKILL.md 正文
Cloud Incident Response Architect
An incident is a coordination problem under uncertainty. The technical fix is usually simple; the reason it takes four hours is unclear roles and missing information.
When to use
- Designing on-call, alerting, and escalation.
- A security incident needs containment without destroying evidence.
- Post-incident review after a major outage.
When NOT to use
- A known failure with an existing runbook — just execute it.
- Pre-launch risk planning →
reliability-architect.
Decision 1 — Roles
| Role | Owns |
|---|---|
| incident commander | decisions, priorities, declaring severity |
| communications lead | status page, stakeholders, updates at fixed cadence |
| operations lead | the actual mitigation work |
| scribe | timeline, decisions, hypotheses |
| subject experts | diagnosis in their area |
- One person can hold several roles in small teams — but the commander must not be the person typing commands.
- Handover protocol for long incidents: written summary, not a verbal chat.
- Relieving an exhausted responder is a decision, not a failure.
Decision 2 — Severity and declaration
- Severity by user impact, defined precisely: S1 total outage, S2 major degradation, S3 partial, S4 internal.
- Anyone can declare; declaring is never punished. Waiting for permission is the expensive mistake.
- Page on symptoms, not causes; the commander decides severity, not the alert.
- Status page updates on a fixed cadence even when there is nothing new.
Decision 3 — Containment
- Stop the bleeding before root cause: roll back, shift traffic, disable the feature flag, scale out.
- Security incidents: revoke credentials first, then investigate. Preserve logs and snapshots before cleanup.
- Isolate without destroying evidence: snapshot the instance, capture the process list, keep the network flow logs.
- Know in advance what you can turn off, and test those switches
(
chaos-engineering-architectgame days).
Decision 4 — Information
- One channel per incident; a shared timeline document that everyone can write to.
- Timestamps in one timezone, one format.
- Dashboards annotated at declaration so the timeline survives after the fact.
- "What changed recently?" is the first question: deploys, config, traffic, certificates, provider status.
Decision 5 — Resolution and review
- Resolution = SLO restored and verified, not "it stopped paging".
- Post-incident review within five days, blameless, focused on system conditions.
- Outputs: timeline, contributing factors, what worked, action items with owners and dates.
- Action items are design changes (timeouts, alerts, runbooks, tests); a doc-only outcome means the review failed.
- Track action completion; publish it — that is how trust is rebuilt.
Decision 6 — Preparedness
- Runbooks for the top 10 failure modes, each tested in a game day.
- On-call rotation with escalation and a documented response-time expectation.
- Tooling tested: paging, status page credentials, break-glass access.
- Practice a provider-wide outage scenario — many systems share one dependency.
Failure modes at scale
- Nobody declares for 40 minutes because everyone assumed someone else would.
- Commander debugging instead of coordinating; no updates for two hours.
- Rollback attempted without knowing what changed; two more changes layered on top.
- Security incident: instance terminated before snapshotting.
- Post-incident actions assigned to nobody, all still open a quarter later.
Verification receipts
- Severity matrix with examples and declaration authority.
- One incident replayed through the timeline: detection, declaration, mitigation.
- Containment switches tested within the last quarter.
- Post-incident actions tracked to closure with owners.
- On-call contacts and escalation verified by a test page.
04 同仓库更多 Skills
interaction-patterns-architect
Architect-level interaction design in code — choosing established patterns over invented ones, feedback and latency behaviour, destructive-action safeguards, navigation and overlay conventions, and the state model behind every interactive surface. Use when a screen has to behave, not just look.
design-system-ops-architect
Architect-level operation of a design system in code — token pipelines and versioning, component API and variant policy, the catalogue as the contract, adoption metrics, and the governance that stops a system from forking into three. Use when a component library must survive several teams.
design-handoff-architect
Architect-level design handoff — what must be specified before implementation starts, tokens over raw values, responsive and content-limit contracts, state matrices, and the review loop that catches drift between design and code. Use when a design is about to become code.
content-microcopy-architect
Architect-level UI content — where strings live, voice and terminology consistency, error and empty-state copy as part of the interface, translation-ready authoring, and the content model a product needs before it multiplies screens. Use when the words in the interface matter as much as the pixels.
ui-ux-implementation-architect
Architect-level direction for turning a design into working UI — the handoff contract, component ownership versus screen-level layout, interaction and state conventions, design QA gates, and the review habits that keep an implementation faithful without frozen pixels. Use before any specific UI implementation skill.
test-strategy-architect
Architect-level test strategy — allocating assertions across unit, integration, contract and end-to-end layers, designing for testability through dependency boundaries, fixture and data strategy, and the flake and coverage policies that keep a suite trustworthy. Use when defining how a codebase is tested.