Skip to content
FortiGate security audit

Audit the live FortiGate configuration behind every policy.

Retrievy collects the live FortiGate configuration over a read-only SSH session and analyzes administration-plane controls, firewall policy hygiene, referenced security profiles, Layer 7 inspection relationships, and change between snapshots. It shows configuration evidence and remediation guidance without pushing changes to the firewall.

The live-device gap

A policy can look complete while its controls remain ineffective.

A firewall rule may reference antivirus, IPS, and web filtering while those engines have reduced or no encrypted-payload visibility because deep inspection is absent or weakened. A permissive rule can also make a later, narrower rule likely unreachable even though both look valid in isolation.

Screenshots and intended templates do not establish what the collected device configuration contains now. Retrievy reads the live configuration and compares it with the prior live snapshot. This page does not claim a FortiManager connector or intended-versus-deployed comparison.

The audit connects management-plane checks, policy relationships, security-profile configuration, evidence, remediation ownership, and the next scan so operators can address the control that causes the downstream exposure.

Configuration and policy audit

Review the administration plane and the traffic-control chain together.

Retrievy analyzes the configuration relationships that determine whether a FortiGate policy is hardened and whether referenced controls can inspect its traffic.

Read-only live collection

The Windows collector connects over SSH with a read-only administrator profile, runs the supported configuration collection, and sends the snapshot for server-side analysis.

Administration-plane hardening

Review trusted hosts, secure administrative access, password and lockout policy, management exposure, session controls, SNMP, logging, and related supported checks.

Policy and Layer 7 analysis

Evaluate policy breadth, logging, referenced objects, likely shadowing, and SSL, antivirus, IPS, application control, web, DNS, and other supported profile configuration.

Evidence, drift, and verification

Retain policy and profile evidence, compare consecutive live snapshots, guide the correction, and use a later scan to verify that the configuration finding cleared.

What Retrievy examines

Configuration evidence from administrator access to downstream inspection.

Some hygiene signals have explicit data requirements. Retrievy qualifies those signals rather than treating missing telemetry as proof of safety.

Check 01

Administrative access

Trusted-host restrictions, default or risky administrators, secure management protocols, password policy, retry and lockout settings, and exposed management services.

Check 02

Session and service hardening

Supported administration-session controls, HTTPS/TLS posture, SSH access, SNMP versions and trusted hosts, logging, time, and other system settings.

Check 03

Firewall policy hygiene

Overly broad sources, destinations, or services, missing logging, disabled rules, and heuristic evidence of policies likely shadowed by earlier rules.

Check 04

SSL inspection

Whether a policy references deep, certificate-only, disabled, missing, or weakened SSL inspection and how that affects visibility into encrypted traffic.

Check 05

Layer 7 profiles

Referenced antivirus, IPS, application-control, web-filter, DNS-filter, and supported related profile settings, including monitor-versus-block behavior where collected.

Check 06

Domino Effect

Configuration-based propagation showing how an SSL inspection gap reduces visibility for downstream controls on encrypted traffic.

Check 07

Objects and dependencies

Referenced address and service objects plus the policy dependencies that provide context for the likely impact of a configuration change.

Check 08

Usage telemetry when enabled

Zero-hit and stale-policy intelligence only when the separate REST policy telemetry needed for hit-count analysis is configured. SSH configuration alone does not prove non-use.

Retrievy FortiGate Policy X-Ray Action Board showing policy risk and Layer 7 control dependencies
FortiGate Policy X-Ray Action Board. Example product data is illustrative.
Product workflow

Audit the live snapshot, fix the root control, verify the next one.

  1. 1 Detect

    Collect without write access

    A read-only SSH profile supplies the live device configuration. Optional policy telemetry adds hit-count context where configured.

  2. 2 Prioritize

    Rank root causes and likely impact

    Policy breadth, administration exposure, referenced profiles, control dependencies, and hygiene evidence help move the root configuration gap ahead of its symptoms.

  3. 3 Remediate

    Use device-specific evidence

    The work item retains the policy, profile, setting, reason, and guidance so the network team can review the production impact and implement the change itself.

  4. 4 Verify

    Re-audit the live device

    A subsequent collection evaluates the live configuration again. Cleared evidence verifies the correction; recurring or new evidence stays visible.

Illustrative example · fictional environment

Security profiles are attached, but HTTPS remains opaque

In this fictional lab, outbound policy 37 references antivirus, IPS, and web-filter profiles but uses certificate-only SSL inspection. The policy permits broad HTTPS traffic from a user network.

Policy
37 · Users-to-Internet
SSL inspection
certificate-inspection
Attached controls
Antivirus, IPS, Web Filter
Audit interpretation
Downstream controls have reduced payload visibility on encrypted traffic

Analysis

The audit does not merely count attached profiles. Policy X-Ray identifies the SSL layer as the shared configuration dependency that limits what the downstream controls can inspect.

Remediation

The firewall owner reviews certificate distribution and required exceptions, then applies an appropriate deep-inspection design through the normal FortiGate change process. Retrievy does not push the change.

Verification

The next live snapshot is analyzed again. The finding clears only when the collected policy and SSL-profile evidence show the corrected configuration.

Why the audit is different

Configuration relationships, with honest telemetry boundaries.

Live snapshot focus

The audit describes the configuration collected from the device, not an unverified intended template or a screenshot supplied months earlier.

Root-cause context

Domino Effect connects an inspection-layer gap to the downstream controls whose visibility it reduces.

Qualified hygiene signals

Likely shadowing is described as heuristic, and unused or stale policy claims require the separate policy telemetry that supports them.

Read-only by design

Retrievy gathers and analyzes evidence. Production firewall changes remain in the network team’s controlled workflow.

Related audit questions

Questions the same snapshot can help answer.

  • Which administrators lack trusted-host restrictions?
  • Which interfaces expose insecure or unnecessary management services?
  • Which policies are overly broad or lack traffic logging?
  • Which later policies are likely shadowed by earlier matching rules?
  • Where does SSL inspection reduce downstream Layer 7 visibility?
  • Which live policy or profile settings changed since the prior snapshot?
Frequently asked

FortiGate Security Audit, answered.

1. Does Retrievy need write access to the FortiGate?
No. The supported collector uses an SSH administrator profile with read-only access for configuration collection. Retrievy analyzes the result and does not push remediation changes.
2. Does the audit compare FortiManager intent with the live device?
No FortiManager intended-state connector is substantiated in the current product. Retrievy analyzes a collected live snapshot and compares it with the previous live snapshot for supported drift.
3. Can a read-only SSH snapshot identify unused policies?
Not by itself. Zero-hit and stale-policy analysis requires separate REST policy telemetry. Configuration-only analysis can still identify disabled, overly broad, missing-logging, and likely shadowed policies.
4. Does “shadowed policy” mean mathematically unreachable in every case?
No. The current shadowing signal is heuristic and uses collected match, service, order, and schedule context. Operators should review the evidence before removing a production rule.
5. How is a FortiGate remediation verified?
A later read-only collection evaluates the live configuration again. The finding moves to verified resolution only when its supporting configuration evidence clears.
Next step

Audit what the firewall is configured to do now.

Connect a FortiGate with read-only access and review management hardening, policy relationships, Layer 7 evidence, and drift.