Question 630
A suspicious launch policy at North Finance has the correct severity, time, and organization but triggers for unrelated detection rules. What policy criterion can narrow the action to its intended detections?
- A blanket increase of every rule’s aggregation threshold
- The alphabetical order of incident titles
- The number of data collectors in the organization
- The selected Rules list in the Automation Policy
Explanation
Correct Answer
D:Automation Policies can select specific generating rules. Limiting the set to the suspicious launch detection prevents actions caused by unrelated incident types.
Incorrect Answers
A:In Automation Policy workflow at North Finance: This weakens detections across the tenant and does not correctly configure the notification scope.
B:For Automation Policy workflow at North Finance: Alert-title sorting does not determine which source rule satisfies an Automation Policy.
C:Considering Automation Policy workflow at North Finance: Distributed collection architecture does not replace selecting which correlation rules should invoke the automation.
Question 631
East Retail wants the correlated incident for failed authentication to include the original User, Source IP, and other event context for investigation. Which rule configuration maps the evidence into the generated alert?
- Only the incident email template without event mapping
- Incident Attributes and Triggered Attributes from rule subpatterns
- Only the global source data retention setting
- Only the rules folder name
Explanation
Correct Answer
B:For correlation-rule investigation at East Retail: The incident definition maps selected filter, group and aggregate attributes into incident fields. Including User and Source IP preserves actionable context for analysts.
Incorrect Answers
A:In correlation-rule investigation at East Retail: An email template cannot recover omitted triggered values unless the incident definition first includes the relevant User and Source IP context.
C:Considering correlation-rule investigation at East Retail: Retention controls availability of evidence but does not decide the fields copied into each generated incident.
D:Under correlation-rule investigation at East Retail: The folder identifies where the rule is stored, not which event attributes the resulting incident retains.
Question 632
After investigating a quarantine alert from email security events at Corporate HQ, the analyst confirms a harmless authorized action and manually closes the alert with a reason. What resolution should be recorded?
- None with Active status indefinitely
- True Positive with Active status after marking it harmless
- False Positive with Manually Cleared status
- True Positive with System Cleared status
Explanation
Correct Answer
C:The event matched detection logic but did not represent the targeted threat. Recording False Positive and manually clearing documents the verified benign quarantine cause.
Incorrect Answers
A:Leaving the incident unresolved obscures the completed triage decision and keeps the reviewed quarantine alert open.
B:For historical or real-time event analysis at Corporate HQ using email security events: A true-positive verdict contradicts the documented authorized behavior; active state also fails to reflect closure.
D:This claims both a real security issue and automatic recovery, while the Corporate HQ investigation found benign activity and a person cleared it.
Question 633
West Operations / host-73: an anomaly job needs to analyze ongoing numeric behavior. Which THREE steps form a defensible training-to-operation pipeline? Choose THREE.
- Run inference on compatible new observations and monitor model drift
- Build a correctly typed numeric feature dataset from the intended event population
- Delete the trained model after every observation without a retraining plan
- Train and evaluate a model on representative historical behavior
- Replace model inputs with a notification recipient list
- Treat all unlabeled records as confirmed attacks
Explanation
Correct Answers
A:Production inference applies the learned representation to later data; monitoring guards against stale assumptions.
B:Training requires coherent numeric measurements rather than missing or unnormalized string values.
D:Model fitting establishes an expected baseline, while evaluation checks whether its scoring is usable for the deployment.
Incorrect Answers
C:Unmanaged deletion prevents stable inference and makes historical scoring impossible to reproduce.
E:Recipient configuration is an alert-delivery mechanism, not a numeric feature set.
F:Anomaly detection learns from unlabeled behavior and is not improved by fabricated universal malicious classifications.
Question 634
Cloud Services wants an active permission failure incident on host-109 to clear after the original rule has not triggered for 60 minutes. The rule is a real-time rule. Which clear-condition option fits?
- Exception effective for 60 minutes starting now
- Original Rule Does Not Trigger with a 60-minute clear period
- Increase notification frequency to one message per minute
- Change the incident title to Resolved but leave status Active
Explanation
Correct Answer
B:For rule deployment or exception at Cloud Services for host-109: The clear timer waits for the specified duration without another rule match, then changes the incident state when the condition is satisfied.
Incorrect Answers
A:An exception suppresses future matching incidents; it is not equivalent to clearing an already-active host-109 alert after inactivity.
C:Considering rule deployment or exception at Cloud Services for host-109: Notification cadence affects delivery only; it cannot use quiet time to clear the incident.
D:Renaming an incident is not a status transition and provides no test of whether permission failure stopped.
Question 635
During an approved 20-minute window, host-162 at Logistics is deliberately generating sign-in events. The existing rule must continue to detect the same behavior on other hosts. What should the analyst configure?
- Clear all previously recorded incidents as the only action
- Raise the count threshold for all hosts permanently
- A scoped, time-bounded exception for the rule
- Disable the rule for every organization until next week
Explanation
Correct Answer
C:The exception can match host-162 and the 20-minute approval window, suppressing only the anticipated incident without silencing other devices.
Incorrect Answers
A:In rule deployment or exception at Logistics for host-162: Clearing historical incidents does not prevent newly matching maintenance events from producing fresh alerts during the window.
B:For rule deployment or exception at Logistics for host-162: A threshold increase changes the detector everywhere and may miss attacks on otherwise quiet devices after the maintenance period.
D:Global disablement creates an avoidable monitoring gap well beyond host-162 and the approved interval.
Question 636
For the response to an incident affecting host-164, corporate HQ has the FortiEMS server in CMDB, but a suspicious launch remediation cannot authenticate to the target. Which FortiSIEM configuration is the relevant prerequisite?
- A supported Enforce On device access credential under Admin > Setup > Credentials
- Change only the rule event filter without validating script execution and target access
- An unrelated notification subject line
- A larger search result-row limit
Explanation
Correct Answer
A:In inventory or lookup query at the SOC for host-164: Remediation uses an approved access method and stored credential to connect to the target device. A CMDB entry alone cannot authorize the reset a controlled service operation.
Incorrect Answers
B:For inventory or lookup query at the SOC for host-164: A detection filter does not establish the target CMDB object, approved credentials or a management route from Run On to Enforce On.
C:Considering inventory or lookup query at the SOC for host-164: Email content does not establish a supported management login for the remediation target.
D:Under inventory or lookup query at the SOC for host-164: Analytics limits affect displayed evidence and cannot provide credentials for the FortiEMS server action.
Question 637
A Research rule must add a Host address from a stopped service incident to a watchlist whose declared element type is IP. The incident attribute is a string label instead. What should be verified?
- The watchlist type matches the type of the incident attribute written to it
- The historical search uses a 30-day default range
- The rule incident title includes the word IP
- The rule has at least three different email recipients
Explanation
Correct Answer
A:In correlation-rule investigation at Research: FortiSIEM rejects incompatible watchlist/incident attribute types. An IP watchlist expects a compatible IP-type incident value, not an unrelated string label.
Incorrect Answers
B:For correlation-rule investigation at Research: Search time horizon does not convert a string incident attribute to an IP-type watchlist entry.
C:Considering correlation-rule investigation at Research: Free-text title labels cannot change the underlying attribute type used for watchlist updates.
D:Under correlation-rule investigation at Research: Email fan-out has no effect on the type compatibility of a Host entry stored in the IP watchlist.
Question 638
Partner Services requires a help-desk ticket or external integration for deny incidents instead of only email. Which Automation Policy capability addresses external workflow?
- Reduce the raw log size to fit an email attachment
- Select a supported integration, ticketing or HTTP/SNMP notification action
- Rename the impacted device to match the help-desk queue
- Add a search display field named ticket
Explanation
Correct Answer
B:FortiSIEM policies can invoke integration policies and supported notification channels, allowing the incident to enter the response system used by Partner Services.
Incorrect Answers
A:In Automation Policy workflow at Partner Services: External action selection does not depend on truncating the source gateway denies record as the mechanism of ticket creation.
C:Considering Automation Policy workflow at Partner Services: Device names are contextual identifiers, not a substitute for configuring the requested external integration action.
D:Under Automation Policy workflow at Partner Services: An Analytics column does not contact a ticket service or deliver the deny incident outside the search result.
Question 639
North Finance wants to tag incidents produced by a connection correlation rule so that analysts can filter the Incident dashboard by this classification. Which step is appropriate?
- Rename the event parser so the rule inherits the parser’s name as a tag
- Create a search result column named tag but leave the rule unchanged
- Select an existing rule tag in Step 3: Define Actions
- Change every incident’s resolution to false positive
Explanation
Correct Answer
C:For North Finance wants to tag incidents produced by a connection correlation rule so that analysts can filter the Incident: Considering rule deployment or exception at North Finance: FortiSIEM lets an analyst attach defined tags to a rule; those tags can appear for incidents in applicable views and support focused triage.
Incorrect Answers
A:For North Finance wants to tag incidents produced by a connection correlation rule so that analysts can filter the Incident: In rule deployment or exception at North Finance: Parser names describe event normalization and are not the same resource as configurable tags attached to rules.
B:For rule deployment or exception at North Finance: A temporary display column does not assign the intended tag to future incidents generated by the rule.
D:Resolution records investigation outcome, not the shared connection classification used to filter incidents.
Question 640
Before FortiSIEM can evaluate whether host-165 is exhibiting an abnormal daily unique destination counts pattern at Research, what is the purpose of a machine-learning training run?
- Convert a CMDB field into a notification transport
- Immediately apply a FortiGate deny policy without a model
- Learn parameters and behavioral patterns from representative historical data
- Rewrite all existing incident resolutions to True Positive
Explanation
Correct Answer
C:Training builds a model using historical daily unique destination counts evidence. Inference later applies that learned model to new observations and can produce anomaly scores or classifications.
Incorrect Answers
A:CMDB stores asset context and does not replace the statistical learning process for the daily unique destination counts input.
B:For machine-learning analysis at Research for host-165: Enforcement is a separate action requiring an incident or policy path; it is not the learning phase of an ML job.
D:Under machine-learning analysis at Research for host-165: Training should not change analyst dispositions on past cases. Past resolutions may be data context, not universal outcome replacements.