Join our Newsletter — 33% off our NHI Course
Home› FAQ› Governance, Ownership & Risk› How should security teams add approval gates to…
Governance, Ownership & Risk

How should security teams add approval gates to infrastructure changes without slowing delivery too much?

← Back to all FAQ
By NHI Mgmt Group Editorial Team Updated October 30, 2026 Domain: Governance, Ownership & Risk

Security teams should place approval gates at the point where infrastructure changes become executable, not after deployment. Use namespace or stack level policies, keep the number of approvers aligned to change risk, and reserve stricter controls for production. The goal is to stop unsafe configuration from reaching apply time while preserving a predictable release flow.

Where approval gates belong in an infrastructure delivery flow

The most effective approval gate sits where a change turns from intent into an executable action, such as a pipeline promotion, policy check, or apply step. That placement lets security review the exact configuration that will run, rather than a stale ticket or design note, and it reduces the chance that a later manual step bypasses the control you thought you had.

For that reason, teams usually get better results by gating at namespace, stack, account, or environment boundaries instead of trying to approve every low-level resource change individually. The control should be close enough to the deployment path to block unsafe infrastructure state, but not so granular that it becomes a bottleneck for routine, low-risk changes.

Approval depth should scale with blast radius. A small, reversible change in a non-production environment should not need the same friction as a production network rule, secret reference, or privileged platform setting. If the approval model does not distinguish between those cases, teams either over-approve everything or create a shadow process that people work around.

How to keep the control useful instead of slow

Good approval design treats risk as the selector for friction, not as a blanket reason to slow all delivery. Pre-approved patterns, policy as code, and environment-specific rules allow most safe changes to pass automatically while forcing human review only when the change crosses a meaningful threshold, such as production exposure, privilege expansion, or deviation from a known-good template.

That approach works best when the approval criterion is objective and machine-checkable. If a change matches a standard module, known configuration profile, or allowed parameter range, it can flow. If it changes scope, expands access, or alters a sensitive dependency, the request should pause for review. This is faster than arguing over every change and more reliable than trying to infer risk from ticket text alone.

Approval also has to align with operational ownership. The team approving the change should understand the platform or workload being changed, while security defines the guardrails and escalation rules. When approvers are disconnected from the execution context, they either become rubber stamps or block changes they cannot confidently assess.

For a practical implementation pattern, NHIMG’s Change Management Policy Template is a useful model for separating routine, normal, and emergency change handling without turning every request into a committee decision.

What breaks approval gates in practice

The most common failure is placing approval after deployment, when the change has already affected the target system. At that point the gate no longer prevents unsafe configuration, it only documents that the unsafe change occurred. A second failure is making approvals too broad, so a single reviewer must sign off on every insignificant change and every high-risk change alike.

Another common problem is relying on ticket approval alone while leaving the pipeline free to apply whatever arrives. That creates a split between governance and execution, and it is easy for teams to drift into “approved in principle” rather than “approved in the exact form that will be applied.” The more the delivery path is automated, the more the approval must be attached to the same automated path.

Infrastructure as code adds a further constraint: the approval should validate the final rendered change, not just the source file. Drift, module reuse, inherited defaults, and environment overlays can produce a materially different outcome from what the initial author saw. NHIMG’s Infrastructure as Code Security Guide is a good reference for thinking about drift and pipeline controls together, rather than as separate problems.

Risk and Threat Considerations

Approval gates reduce the chance that unsafe infrastructure reaches production, but they can also create blind spots if the workflow does not stop the exact executable change. The risk is not only accidental misconfiguration, it is also policy bypass through alternative paths, emergency exceptions, or overbroad approver trust.

Failure mechanism: If approvals are attached to tickets, chat messages, or late-stage deployment reviews instead of the actual apply action, an unsafe configuration can still be executed through the pipeline. If approvers are too far removed from the change context, they may approve without understanding the blast radius.

Impact: The organisation can end up with production drift, privilege expansion, service disruption, or security exposure that appears “approved” in governance records but was never truly controlled at execution time.

Standards & Framework Alignment

This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.

NIST SP 800-53 Rev 5 and CIS Controls v8 set the technical controls, while ISO/IEC 27001:2022 defines the regulatory obligations.

FrameworkControl / ReferenceRelevance
NIST SP 800-53 Rev 5CM-3 — Configuration Change ControlInfrastructure approvals are change control over executable system state.
CM-6 — Configuration SettingsThe topic centers on controlling approved configuration state, especially for production.
AC-6 — Least PrivilegeApproval depth should scale with privilege and blast radius.
Recommendation — Require approval before applying infrastructure changes and document the authorized change path. Define approved baseline settings and block deviations from those settings in production. Limit who can approve and execute high-impact infrastructure changes.
ISO/IEC 27001:2022A.8.32 — Change managementThe question is directly about governing infrastructure change without disrupting delivery.
Recommendation — Apply controlled change approval and testing before production deployment.
CIS Controls v8CIS-4 — Secure Configuration of Enterprise Assets and SoftwareApproval gates protect approved configuration from unsafe drift and misuse.
Recommendation — Enforce secure configuration baselines and require review for risky deviations.

Practitioner Guidance

What to prioritise: Put the gate at the last controllable point before infrastructure state changes, then reserve manual review for changes that materially increase blast radius, privilege, or production exposure.

What to verify: Confirm that the approval decision is bound to the exact plan, manifest, or rendered change that will be applied, and that there is no alternate path around the gate through direct console edits or emergency automation.

Decision rule: If the change is low-risk and matches a standard pattern, let policy automate the approval path; if it alters production scope, access, or shared platform controls, require human sign-off before apply.

Practitioner takeaway: The best approval gate is narrow enough to protect the execution point, but smart enough to exempt routine change, otherwise it becomes either a bottleneck or a paper control.

Free weekly newsletter

Subscribe to the NHI & AI Identity Journal

The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.

Bonus 33% off our NHI Course when you subscribe.

NHIMG Editorial Note
Reviewed and updated by the NHIMG editorial team on October 30, 2026.
NHI Mgmt Group — the #1 independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org