A.8.9
Configuration management
ISO 27001 · ISO/IEC 27001:2022 · Last verified July 2026
Objective
Establish, document, implement, monitor, and review configurations — including security baselines — for hardware, software, services, and networks.
Points of focus
- Security baselines defined for cloud accounts, container images, and OS builds
- Configuration changes go through version control and peer review, not console clicks
- Drift between declared and running state is detected automatically
- Baselines are updated when new services or regions are onboarded
- Emergency out-of-band changes are logged and reconciled back into IaC
Implementation notes
Treat every production surface — cloud accounts, Kubernetes clusters, VM images, network ACLs, and SaaS platform settings (e.g. GitHub org policies, Okta tenant config) — as code with a single source of truth. Build golden images and base container layers from a documented hardening checklist (CIS or your own distilled version) and rebuild them on a fixed cadence rather than patching running instances by hand. Wire drift detection into the same alerting pipeline used for security incidents so unauthorized changes get triaged, not archived in a dashboard. When engineers must make an emergency console change, require a same-week follow-up pull request that reconciles the state; track these as a named exception category so the volume itself becomes a signal at management review. Extend the same discipline to SaaS-of-SaaS configuration — identity provider policies, CI/CD runner permissions, and data-warehouse access grants — since these are frequently the systems attackers target first and the ones most often left to manual, undocumented tweaks.
Audit tip: Bring up a real Terraform module for a production resource, show the linked pull request with an approving reviewer, and then show the drift-detection dashboard covering the same resource for the last 90 days — that three-part chain is far stronger than a written policy alone.
Evidence auditors typically request:
- Terraform/CloudFormation modules with a CODEOWNERS-approved baseline for prod accounts
- CIS Benchmark or internal hardening checklist applied to golden AMIs or base container images
- Drift-detection tool output (AWS Config, driftctl, or Terraform plan diffs) from the last quarter
- Kubernetes admission-controller policy (OPA/Gatekeeper or Kyverno) blocking non-compliant pod specs
- Pull request history showing reviewer approval for a sampled infrastructure change
Common gaps
- Console changes made directly in AWS/GCP during an incident are never back-ported into Terraform, so the repo silently diverges from production
- A hardening checklist exists for the base AMI but new services get spun up from unpatched community images
- Drift-detection alerts fire into a channel nobody monitors, so weeks-old unauthorized changes go unnoticed
Cross-Framework Mapping
| Framework | Requirement | Implementation note |
|---|---|---|
| ISO 27001 | A.8.9 | This control |
| SOC 2 | CC8.1 | Related Trust Services Criteria themes — map in your crosswalk; not identical requirements. |
Primary sources
- ISO/IEC 27001:2022 Annex A: ISO/IEC 27001:2022 Annex A (A.8.9)