Fortinet FCP_FSM_AN-7.2 (FCP - FortiSIEM 7.2 Analyst) Practice Test 6

Question 50

At Research, an anomalous daily unique destination counts model was trained on a field imported as text, so numeric computations fail. Which correction most directly addresses the ML input?

  1. Mark all anomalous records as resolved before training
  2. Use a numeric report field for the measure being modeled
  3. Treat a normalized event query as a completed ML training model
  4. Replace the field with the incident title

Explanation

Correct Answer

B:

Numeric anomaly models rely on continuous or count-like features. Correctly typed daily unique destination counts data allows the learner to evaluate ranges and deviations.

Incorrect Answers

A:

Incident resolution is not a substitute for correcting the underlying daily unique destination counts feature type.

C:

Considering machine-learning analysis at Research: An event search selects records, while a model must be trained on the intended dataset and then applied to compatible new features.

D:

Titles are labels for people and do not provide the numeric measurements needed to learn variation in daily unique destination counts.

Question 51

Cloud Services knows a legitimate scheduled scanner will produce policy violation events on host-128 during a maintenance window. The detection should remain active for all other assets and times. Which rule feature targets the exception?

  1. Change the incident severity to low with no exception
  2. A rule exception constrained by relevant attributes and an effective period
  3. Delete the parser for the log source
  4. Disable the entire rule permanently

Explanation

Correct Answer

B:

Rule exceptions suppress incidents for matching host-128 conditions during a defined single or recurring interval, preserving detection elsewhere.

Incorrect Answers

A:

In correlation-rule investigation at Cloud Services for host-128: Lowering priority still creates incidents for the expected scan and does not delimit the approved maintenance period.

C:

Removing event parsing would harm all observations from firewall session logs and is not a precise, time-bound exception.

D:

Global disablement discards all policy violation detections beyond the approved host-128 maintenance exception.

Question 52

In East Retail, a stored search matches 5 blocked lookup records from DNS resolution records, including only 3 unique Queried Domain values for host-105. To measure unique targets rather than raw volume, which aggregate is required?

  1. AVG(Queried Domain)
  2. SUM(Queried Domain)
  3. COUNT(DISTINCT Queried Domain)
  4. COUNT(Matched Events)

Explanation

Correct Answer

C:

This removes repeated Queried Domain values before counting. For host-105, the expected unique-target measure is 3 rather than the 5 event rows.

Incorrect Answers

A:

In aggregate-query calculation at East Retail for host-105 using DNS resolution records: An average of an identifier would not quantify unique Queried Domain values, even if the attribute happened to be encoded numerically.

B:

For aggregate-query calculation at East Retail for host-105 using DNS resolution records: SUM is intended for numeric measures and is not a meaningful distinct-count function for the Queried Domain identity field in these DNS resolution records records.

D:

Under aggregate-query calculation at East Retail for host-105 using DNS resolution records: This produces 5 because each qualifying event contributes one row. It cannot distinguish repeat connections to a target from a new Queried Domain.

Question 53

A Research analyst wants the event’s Queried Domain and Source IP values visible in a generated incident but notices only the incident category appears. Which editor area should be reviewed?

  1. Modify the case label without reviewing rule state or exceptions
  2. Define Action and the incident’s Triggered Attributes mapping
  3. The lookup-table schema editor for unrelated enrichment
  4. The FortiGate DHCP lease duration

Explanation

Correct Answer

B:

Triggered Attributes controls which details from matching events are carried into the incident for dashboards and analysis, including Source IP and Queried Domain.

Incorrect Answers

A:

In rule deployment or exception at Research: Case metadata does not activate an inactive rule or correctly scope a rule exception for future event matches.

C:

Considering rule deployment or exception at Research: Lookup tables can supply context, but the missing direct event fields are selected in the incident definition.

D:

For A Research analyst wants the event’s Queried Domain and Source IP values visible in a generated incident but notices only: Under rule deployment or exception at Research: DHCP duration does not determine which parsed attributes are copied into the FortiSIEM incident.

Question 54

For a connection incident in Cloud Services, the analyst wants to record findings that a responder can read without falsely treating the alert as resolved. Which incident operation is suitable?

  1. Delete the device from CMDB so the incident loses context
  2. Add an incident comment or remediation advice while preserving current status
  3. Set False Positive and clear every incident sharing the category
  4. Raise the global rule count threshold as a substitute for notes

Explanation

Correct Answer

B:

FortiSIEM supports comments on incident records and remediation notes in rules. Documentation helps responders act on host-207 without prematurely clearing the live alert.

Incorrect Answers

A:

In incident-management case at Cloud Services: Removing an inventory object discards relevant asset information and does not constitute safe analyst documentation.

C:

This asserts a benign verdict and changes lifecycle state for alerts that may still represent genuine connection activity.

D:

Under incident-management case at Cloud Services: Changing detection conditions alters future alerts, not the investigation narrative that a responder needs now.

Question 55

At Corporate HQ, a built-in FortiSIEM ML job template is visible for nightly database write rate. The team assumes the template alone will produce valid scores. What must occur first?

  1. Delete the historical event database
  2. Configure the relevant data source and train a usable model
  3. Rename the template to the incident severity value
  4. Create an unrelated ZTNA firewall policy

Explanation

Correct Answer

B:

A job template provides a starting definition, not a trained statistical baseline. The Corporate HQ dataset and training results must support subsequent inference.

Incorrect Answers

A:

In machine-learning analysis at Corporate HQ: Historical observations are normally necessary for establishing the behavior the ML job will learn.

C:

A cosmetic name change neither supplies nightly database write rate samples nor fits model parameters.

D:

External network enforcement does not build a trained model for the nightly database write rate feature.

Question 56

City Branch has correlated several failed login incidents and related events into one investigation that should have an owner, workflow status, and associated evidence. Which FortiSIEM object is most appropriate?

  1. A notification severity threshold alone
  2. An unrelated CMDB lookup key
  3. A case associated with the relevant incidents/events
  4. A free-text event keyword saved as a report

Explanation

Correct Answer

C:

Case management groups related incidents into an assigned work item with investigation ownership and progression, rather than treating every host-44 alert independently.

Incorrect Answers

A:

In incident-management case at City Branch: A threshold may control alert delivery but cannot connect multiple findings into one assigned investigation.

B:

For incident-management case at City Branch: A lookup key helps enrichment but does not provide an investigator owner and lifecycle for the related incident cluster.

D:

A saved search can surface evidence but is not the managed case workflow with assignment and status for City Branch.

Question 57

For Partner Services the two Host groups contain 5 and 3 failed sign-in events. The analyst needs total event count per Host, without mixing the sources. Which pair of settings is necessary?

  1. Group By User and show AVG(Sent Bytes)
  2. Filter to one Host before counting and treat it as the whole result
  3. Group By Host and show COUNT(Matched Events)
  4. Show only COUNT(Matched Events) without a Group By attribute

Explanation

Correct Answer

C:

Considering aggregate-query calculation at Partner Services with the specified 5 value: Host partitions the matching events into independent groups while COUNT tallies their individual 5 and 3 records.

Incorrect Answers

A:

In aggregate-query calculation at Partner Services with the specified 5 value: This would summarize a different dimension and average traffic volume rather than count failed sign-in events per Host.

B:

For aggregate-query calculation at Partner Services with the specified 5 value: Excluding the other source can produce one count but cannot satisfy the simultaneous side-by-side Host comparison.

D:

Under aggregate-query calculation at Partner Services with the specified 5 value: A global COUNT collapses the 8 events together and loses the separate per-Host quantities.

Question 58

Cloud Services plans to reuse an inner event search that lists Source IP entities for nested correlation. Which preparation makes its values available to the outer query?

  1. Erase the inner report display columns before saving
  2. Save only a screenshot of the inner results
  3. Save an inner report with the intended comparison attribute among its output fields
  4. Create a notification policy for each prior event

Explanation

Correct Answer

C:

Considering nested search comparison at Cloud Services: The outer nested filter references an output column from the saved inner report. Including Source IP makes the membership values explicit and type-checkable.

Incorrect Answers

A:

In nested search comparison at Cloud Services: Without a relevant output attribute, the outer filter lacks a typed list to compare with the current event values.

B:

For nested search comparison at Cloud Services: A screenshot documents findings but cannot provide the live Source IP values that a nested query evaluates.

D:

Under nested search comparison at Cloud Services: Alert routing is independent of saving an executable inner report that returns Source IP values.

Question 59

In East Retail, a stored search matches 9 stopped service records from service health records, including only 7 unique Service values for host-128. To measure unique targets rather than raw volume, which aggregate is required?

  1. AVG(Service)
  2. COUNT(DISTINCT Service)
  3. SUM(Service)
  4. COUNT(Matched Events)

Explanation

Correct Answer

B:

This removes repeated Service values before counting. For host-128, the expected unique-target measure is 7 rather than the 9 event rows.

Incorrect Answers

A:

In aggregate-query calculation at East Retail for host-128 using service health records: An average of an identifier would not quantify unique Service values, even if the attribute happened to be encoded numerically.

C:

Considering aggregate-query calculation at East Retail for host-128 using service health records: SUM is intended for numeric measures and is not a meaningful distinct-count function for the Service identity field in these service health records records.

D:

Under aggregate-query calculation at East Retail for host-128 using service health records: This produces 9 because each qualifying event contributes one row. It cannot distinguish repeat connections to a target from a new Service.

Question 60

An MSSP analyst is authorized to inspect Corporate HQ only. A query for permission failure events from database audit records unexpectedly includes tenant peers. Which search setting should be restricted?

  1. Add a display column showing the event time
  2. Group the mixed result by User
  3. Increase the maximum rows returned by the search
  4. Select the intended organization in the service-provider search scope

Explanation

Correct Answer

D:

FortiSIEM service-provider searches support organization selection. Constraining scope to Corporate HQ prevents the query from including the other tenants’ event sets.

Incorrect Answers

A:

In historical or real-time event analysis at Corporate HQ using database audit records: Event timestamps assist investigation but do not replace the search’s organization selection when isolating one customer.

B:

Grouping by User reorganizes the already selected records and does not enforce the organizational boundary for Corporate HQ.

C:

Row limits affect truncation, not data tenancy; a higher limit would reveal more mixed-tenant events instead of isolating Corporate HQ.

FCP - FortiSIEM 7.2 Analyst (FCP_FSM_AN-7.2) Practice Tests