Fortinet NSE4_FGT_AD-7.6 Practice Test 38
Question 370
A FortiGate web-inspection rollout switches an application policy from flow to proxy mode. During peak downloads the web proxy process and first-byte latency increase sharply, but a required content-filter feature may be proxy-only. What measurements and feature checks should guide tuning or a controlled return to flow inspection?
- Measure proxy WAD CPU and first-byte latency against mandatory inspection features; test a scoped, supported flow-mode alternative only if security behavior remains equivalent.
- Switch every security policy to flow mode immediately after a latency complaint without checking whether any required web-filter action is proxy-only.
- Increase the session timeout for every application.
- Change SD-WAN default-route priority for the affected web application after proxy inspection is enabled, without measuring its SSL processing or first-byte delay.
Explanation
Correct Answer
A:Proxy inspection can buffer and reconstruct content, producing additional latency for some file-heavy applications, but certain required filtering features may rely on its feature set. Correlate WAD/process resource diagnostics with representative downloads and determine whether flow inspection supports each mandatory control. Test a scoped profile change with security and performance measurements rather than globally disabling inspection. In FortiOS 7.6.0, proxy-based inspection is unsupported on physical FortiGate models with 2 GB RAM or less; confirm model capacity before selecting that mode (VM behavior differs).
Incorrect Answers
B:Flow mode can reduce processing in some deployments but may lack required proxy features. Measure WAD costs and test a supported feature-equivalent adjustment.
C:Idle expiration may not cause a per-request inspection delay. Examine CPU, concurrent scans, file sizes and policy mode to locate the actual bottleneck.
D:Route changes may affect network latency but cannot prove a regression rooted in proxy buffers or WAD CPU. Isolate proxy versus path timing first.
Question 371
An FGCP cluster has override disabled and identical configured priorities. After a power cycle, both monitored interfaces are healthy. The former primary restarted 3 minutes ago, while its peer has 48 minutes of HA uptime; the default 300-second HA-uptime difference margin is unchanged. The peer remains primary. Which FortiOS 7.6.0 election evidence explains this result?
- Raise the recently restarted member’s HA device priority and expect immediate takeover although override is disabled and its HA uptime is 45 minutes lower.
- Compare only the hardware system uptime since manufacture instead of HA uptime and the Primary selected using history reported by get system ha status.
- Enable HA session pickup and assume synchronized sessions determine election winner when monitored-link health and configured priorities are otherwise equal.
- Apply the FGCP 300-second uptime comparison: with override disabled and 45-minute difference, the older healthy member wins.
Explanation
Correct Answer
D:With override disabled, FortiOS favors HA uptime over numeric priority after applicable health criteria. The 48-minute versus 3-minute HA uptimes differ by 45 minutes, greater than the documented default 300-second comparison margin, so the longer-running healthy member can stay primary. Verify this result in get system ha status and the primary-selection history; the election algorithm, not a presumed fixed primary role, determines the outcome.
Incorrect Answers
A:With override disabled and a difference above the 300-second threshold, HA operating age can precede priority. Read actual election state before changing config.
B:FGCP uses HA uptime and monitored-interface election criteria rather than simply the ordinary displayed operating-system uptime or an old role label. The default election age margin is 300 seconds; verify current HA uptime and the recorded Primary selected using reason.
C:Session pickup helps preserve eligible sessions when an HA member becomes primary, but synchronized sessions are not a primary-unit selection criterion. This scenario turns on the actual monitored-link and uptime comparison, including the documented election age threshold, not on the count of synchronized sessions.
Question 372
A private application needs UDP/TCP 53 for DNS resolution and TCP 443 for its web interface. A teammate suggests an ALL-service firewall policy. Which definition is preferable?
- Permit TCP 53 and TCP 443 only, relying on DNS TCP fallback to cover all ordinary UDP DNS client queries for the private application.
- Permit UDP 53 and UDP 443 and assume encrypted HTTPS application data uses the same UDP transport as the DNS query.
- Permit every service between the two application subnets and rely on the server's host firewall as the only restriction.
- Permit only DNS UDP/TCP 53 and application HTTPS TCP 443 to their exact authorized servers and sources.
Explanation
Correct Answer
D:An internal application that uses DNS and HTTPS should be authorized by firewall rules matching the correct interfaces, source/destination objects and specifically required service definitions. Avoid an all-services permit as a shortcut; it can open administration and unrelated protocols. Check whether DNS uses a separate resolver path and whether HTTPS inspection or authentication is required. Validate permitted lookups and HTTPS sessions while confirming a nonapproved port remains blocked.
Incorrect Answers
A:DNS typically uses UDP/53 for common queries, and the case requires both transports. Omitting UDP can break name resolution despite working HTTPS.
B:The specified web interface requires TCP/443; allowing UDP/443 alone does not satisfy it. Restrict to the stated DNS and HTTPS services.
C:That expands lateral reachability far beyond DNS and HTTPS. FortiGate's service objects should constrain transport and ports at the network boundary, then application authentication provides a separate control. An ALL-service rule exposes unnecessary internal protocols; permit DNS and HTTPS traffic only as documented.
Question 373
All SD-WAN member probes show acceptable latency and loss, but users report slow page loads on one SaaS application. FortiGate SSL deep inspection is enabled. What should be compared?
- Treat healthy SD-WAN probe latency and loss as proof the entire SaaS page load meets business response targets without checking DNS, TLS processing or server timing.
- Inspect the existing SaaS SD-WAN performance SLA and raise jitter limits so both providers remain eligible while DNS, TLS and server timings stay unmeasured.
- Correlate actual SaaS response latency with member SLA, DNS and SSL-inspection process timing during the incident; investigate whether the bottleneck is DNS, the remote service or inspected proxy processing.
- Switch the SaaS policy from deep to certificate-only SSL inspection immediately and permanently, attributing all delay to decryption without checking DNS, server response times or actual FortiGate process measurements.
Explanation
Correct Answer
C:Normal aggregate interface counters do not establish the experience of a specific web application. Compare SD-WAN member latency, jitter, loss and SLA status with application response timing, DNS resolution and the actual firewall-session egress path. Test during the reported slow interval and determine whether degradation is link quality, server response or inspection delay. Raising SLA thresholds arbitrarily could move traffic to a worse member while masking the true bottleneck.
Incorrect Answers
A:Performance-SLA probes quantify selected path latency, loss and jitter, but a normal probe does not prove low time-to-first-byte for the encrypted SaaS transaction. The application may still be delayed by DNS, upstream processing or inspection. Compare transaction timing on both paths before blaming the ISP.
B:SLA criteria are useful when the application is suffering from a measurable WAN quality failure, but the available probes already show acceptable latency and loss. Relaxing jitter without recording actual SaaS transaction timing could worsen path quality and cannot attribute slow pages to DNS, remote service or deep inspection. Measure end-to-end application phases alongside member and FortiGate process counters first.
D:Deep inspection may add latency for some applications, but healthy WAN probes alone cannot establish which part of the transaction is slow. A blind certificate-only switch removes content visibility and cannot prove whether SaaS response time, DNS, overloaded security processes or steering caused the delay. Compare timed evidence first and test only scoped changes.
Question 374
A security policy disables upstream certificate verification during outbound deep inspection. An attacker can substitute a lookalike destination certificate. What risk is introduced?
- Restore upstream certificate chain, hostname and validity checks in deep inspection; endpoint trust alone is insufficient.
- Deploy the FortiGate signing CA to every browser and treat trusted substituted certificates as proof the upstream origin certificate was verified.
- Enable traffic logging and assume that the recorded TLS session prevents server impersonation when certificate validation is disabled.
- Disable invalid-origin certificate alerts and depend on user-visible browser padlocks after decryption as the only origin authentication mechanism.
Explanation
Correct Answer
A:Turning off upstream server-certificate validation allows a decryption gateway to trust fraudulent or impersonated destinations despite presenting its own trusted certificate to the client. Review certificate-chain validation, revocation behavior, SNI and exceptions in the matched SSL inspection profile. Test a deliberately untrusted origin in a controlled environment and verify the expected block or warning. Do not confuse endpoint trust of the inspection CA with FortiGate's independent trust decision about the remote server.
Incorrect Answers
B:Endpoint trust of the inspection signer does not authenticate the remote server. Restore upstream issuer, hostname and validity checks in the profile.
C:Logs preserve evidence but cannot establish a cryptographic identity check that was never performed. Require appropriate upstream certificate-chain and hostname validation for the intended service. Logging observes a handshake after the fact; it cannot substitute for authenticating the remote TLS endpoint.
D:A substituted trusted leaf can display a padlock even if upstream validation is disabled. The FortiGate must enforce origin identity validation.
Question 375
An IKEv2 hub and spoke authenticate successfully. The spoke proposes local 10.40.8.0/24 and remote 10.2.0.0/24, but the hub child-SA log shows that its local and remote selectors were entered in the same direction as the spoke instead of being reversed. The log reports no matching child SA. Which configuration should be compared and corrected?
- Compare and reverse the hub-side local/remote protected-network selectors to match the spoke from the hub perspective, then verify both child-SA proposal and policy coverage.
- Review the hub static route to 10.40.8.0/24 and add a more specific IPsec-interface route while preserving the failed child-SA selector configuration.
- Inspect child-SA lifetime and increase the phase-two rekey and DPD tolerances without correcting the logged same-direction hub-side subnet selectors.
- Recreate the IKEv2 phase-one association with a fresh shared secret while keeping the child-SA selector ranges unchanged.
Explanation
Correct Answer
A:Phase-two local and remote subnets are defined from each VPN peer perspective. The spoke local 10.40.8.0/24 must appear as the hub remote protected prefix, while the spoke remote 10.2.0.0/24 is the hub local protected prefix. Align those directions and compatible ESP/PFS settings, then confirm child-SA formation before testing forwarded routes and policy.
Incorrect Answers
B:Routing to a remote protected prefix is necessary for packets after an SA is negotiated, but the documented failure is no matching child SA because the hub selectors are entered from the wrong endpoint perspective. Routing alone cannot make a rejected selector acceptable; reconcile hub/spoke local and remote ranges first.
C:Timers can help with genuine rekey or packet-loss failures, but the hub and spoke already authenticate and the log explicitly identifies incompatible local/remote traffic selectors. The hub must reverse its protected network perspective and verify overlapping child-SA scope; waiting longer cannot change those addresses.
D:IKEv2 phase-one already authenticated successfully, so rotating its credential would not fix the logged child-SA protected-prefix direction mismatch. Repair the hub-side local/remote proposal and validate matching policies.
Question 376
Memory usage is marked high on the FortiGate dashboard, but user traffic is currently working. An operator wants to reboot immediately. What should be checked first?
- Collect FortiOS memory process, session and conserve-mode trends before treating a dashboard alert as a failure; verify current traffic works and conserve thresholds are not crossed before disruptive remediation.
- Inspect the highest-memory FortiOS process and restart that process immediately based on the dashboard percentage alone.
- Review aggregate AV and IPS memory counters and disable both security profiles globally to reduce the dashboard high-memory warning.
- Disable resource alerts globally while existing user traffic works, and assume the high-memory mark can never become a forwarding problem.
Explanation
Correct Answer
A:A dashboard memory warning should be correlated with actual memory usage, conserve-mode state, process consumption and system events before concluding that forwarding is already broken. Capture current diagnostic snapshots and a relevant time history so a transient peak can be distinguished from sustained resource exhaustion. If specific inspection processes are responsible, review their workloads and supported remediation. Disabling all security profiles simply because a graph is yellow would weaken protection without identifying the cause.
Incorrect Answers
B:Process identification is part of a proper resource review, but memory consumption can be expected for active sessions and buffers. Without conserve-mode threshold, growth-rate and session evidence, restarting the apparent top process may interrupt traffic unnecessarily and conceal the cause. Collect trends before intervention.
C:Security profiles can contribute to memory pressure, but the mere high-memory mark does not establish which process is at fault or that conserve thresholds have been crossed. Global disabling strips protection from unaffected traffic. Inspect process and session trends, conserve status and workload before any scoped tuning.
D:An alert alone is not an outage, but ignoring sustained growth can miss approaching conserve mode. Trend actual memory and session state.
Question 377
A firewall traffic log lists an unexpected public source address for an internal client. The policy uses central source NAT. What additional evidence should clarify the log?
- Correlate the logged public source with the internal pre-SNAT client and chosen NAT pool/session translation.
- Use the public post-SNAT address as the original user source in the security incident report without matching the session to the internal client.
- Inspect VIP destination mappings for the public address and create a destination-NAT object to reconcile the unexpected source shown in traffic logs.
- Review Central SNAT pool configuration and disable WAN egress source NAT so future logs show private addresses directly.
Explanation
Correct Answer
A:Unexpected source addresses in logs can be a normal consequence of central or policy-based source NAT. Identify whether the viewed field represents the original initiator, translated source or reply direction, and correlate it with the matching firewall policy, central SNAT or IP pool and session entry. Confirm packet captures on both ingress and egress to establish actual translation. Do not mistake a public SNAT address for a new unauthorized internal client.
Incorrect Answers
B:A post-SNAT public source address can represent many distinct internal users at different times or ports. It cannot be substituted as the original initiating identity. The audit requires the pretranslation client and source port, the translated tuple and the corresponding session timestamp.
C:VIPs translate packet destinations for publishing or destination routing, not the outbound client source chosen by Central SNAT. The log’s public source can be explained only by joining its pre-translation client to the winning Central SNAT rule, IP pool and egress. Do not add a destination mapping.
D:Disabling NAT may expose private addresses to the ISP and break return connectivity when it does not route RFC1918 networks. It also loses the mapping needed for existing sessions. For attribution, correlate original and translated source in the session or log rather than altering a functioning NAT design.
Question 378
A FortiGate DNS Filter blocks malicious domains in ordinary UDP/53 queries, but managed browsers have switched to a third-party DNS-over-HTTPS resolver on TCP/443. The existing firewall policy has no deep inspection or DNS-filter coverage for that encrypted resolver path. Which FortiOS 7.6.0 investigation and policy correction can close the observed gap without indiscriminately blocking business HTTPS?
- Trace the browsers’ actual DoH/DoT resolver sessions, then apply supported full SSL inspection and DNS Filter to authorized encrypted-DNS paths and restrict unapproved resolver access on the matched policy.
- Add UDP/53 DNS-filter blocks for newly identified public DoH providers and assume their browser HTTPS queries are redirected to ordinary DNS.
- Turn off SSL inspection everywhere to guarantee that DNS over HTTPS is visible to the DNS filter.
- Block the approved resolver IP while keeping arbitrary third-party DoH endpoints reachable, expecting client applications to fall back to protected DNS.
Explanation
Correct Answer
A:In FortiOS 7.6.0, DNS Filter can inspect supported DoH and DoT sessions in flow or proxy inspection mode when the effective firewall policy applies compatible full SSL inspection and DNS filtering. Trace the user, selected resolver and actual encrypted protocol, test the category Block verdict and restrict unauthorized encrypted resolver destinations without applying a blanket TCP/443 deny. Plaintext UDP/53 controls do not automatically classify an unrelated encrypted DNS session.
Incorrect Answers
B:A UDP/53 rule applies to conventional plaintext DNS, not arbitrary browser DoH carried as HTTPS on TCP/443. Blocking provider addresses only on the plaintext-DNS port cannot redirect their encrypted resolver requests into the existing DNS Filter. Instead validate which FortiOS 7.6.0 inspected DNS-policy path sees the real DoH/DoT session and apply a narrowly authorized control.
C:Disabling the full SSL inspection capability cannot make encrypted DoH or DoT content visible to a DNS Filter. FortiOS 7.6.0 supports inspected DoT/DoH traffic with appropriate SSL/SSH inspection and a DNS Filter profile on the matching firewall rule; identify and test that path instead of making TLS visibility worse.
D:Blocking only the enterprise-approved resolver while leaving the third-party encrypted resolver path open would not restore trustworthy filtering; it could increase fallback to uninspected DoH. Verify effective policy matching and enforce the approved DNS resolver and compatible DoH/DoT inspection path without interrupting legitimate enterprise queries.
Question 379
Two WANs are in one SD-WAN zone. The FortiGate has default routes, but one member's gateway is not reachable and the member is never selected. Which check matters?
- Inspect the SD-WAN member preference rule and force the unreachable member to be the preferred path during testing.
- Disable the remaining healthy member so the route table must choose the broken link.
- Check the ineligible member’s gateway/ARP, installed route and performance-SLA status while leaving healthy WAN paths active; correct the failed dependency before retesting selection.
- Review the common SD-WAN zone ACCEPT policy and broaden its service to ALL so both WAN gateways can participate.
Explanation
Correct Answer
C:Multiple default routes and an SD-WAN zone do not establish which member can actually forward a given session. Compare the installed routing table, source and destination tuple, SD-WAN service-rule match, member gateway, health-check state and route eligibility. Verify the selected egress using diagnostic output and a test flow, especially when one member lacks a valid gateway. This tests the forwarding decision rather than relying on a configured member list alone.
Incorrect Answers
A:A service preference cannot supply layer-two adjacency to a gateway that fails ARP or has no eligible route. Forcing traffic onto it risks blackholing new sessions. Confirm its on-link gateway, route installation and performance SLA state while keeping healthy members available.
B:Removing the fallback can create a complete outage without repairing the gateway. Validate actual member reachability and configuration first. Disabling the working member creates avoidable loss; keep it available while repairing the ineligible alternate path.
D:The zone security policy is necessary for eligible forwarding but does not establish that an individual member gateway is reachable or included in the route table. A broader service can expand permitted traffic without correcting ARP, next-hop route eligibility or health failures. Diagnose the member before adjusting policies.
Question 380
A compliance report must show whether an individual HTTPS request's content passed through deep inspection. Current traffic logs list the forwarding policy but no SSL profile. Which proof is necessary?
- Correlate applied SSL profile and actual certificate/content events with the HTTPS session before claiming decryption occurred.
- Inspect the HTTPS ACCEPT traffic log for the policy ID and use its accepted byte counter as evidence that the encrypted content was malware-scanned.
- Review the global SSL/SSH inspection feature status and export a screenshot of the configured deep-inspection profile for the audit.
- Use a DNS request log to infer that all later TLS URL paths were decrypted and inspected.
Explanation
Correct Answer
A:An audit must distinguish TLS traffic that was merely seen from traffic actually decrypted or inspected for content. Correlate the session's policy ID with the applied SSL/SSH inspection profile, the chosen inspection mode, certificate events and relevant security-profile logs. Verify an authorized controlled HTTPS test shows the expected content inspection, not only a successful TLS handshake. A certificate-inspection profile alone cannot prove antivirus saw the encrypted response body.
Incorrect Answers
B:Traffic bytes and policy acceptance establish forwarding, not a particular SSL/SSH inspection depth or content-verdict action. The same policy could use certificate-only inspection or a different profile. Correlate the actual SSL profile and decryption/content security logs with this exact request before claiming payload scanning.
C:Global availability of a deep-inspection profile does not prove that a particular HTTPS session matched it or that encrypted payload was decoded. The compliance claim requires connection-specific policy/profile association, applicable certificate evidence and security-event disposition rather than an unused configuration screenshot.
D:DNS may reveal a name, not the HTTPS payload or full URL path. Check which TLS profile processed the session and whether the applicable privacy exemption applied. A DNS event reveals hostname resolution but cannot demonstrate decryption of a later HTTPS request path.