Machine Learning Anomaly Detected

Rule ID

PH_Rule_ML_1

Default Status

Enabled

Description

Detects a machine learning based anomaly

Severity

MITRE ATT&CK® Tactics

Policy Violation

MITRE ATT&CK® Techniques

No Technique Specified

Impacts

Server

Data Source

FortiSIEM Machine Learning based Detecion

Detection

Correlation

Remediation Guidance

No remediation guidance specified

Time Window

If the following pattern or patterns match an ingested event within the given time window in seconds, trigger an incident.

300 seconds

Trigger Conditions

If the following defined pattern/s occur within a 300 second time window.

Anomaly

SubPattern Definitions

SubPattern Name: Anomaly

This is the named definition of the event query, this is important if multiple subpatterns are defined to distinguish them.

SubPattern Query

This is the query logic that matches incoming events

eventType = "PH_ML_ANOMALY_DETECTED"

Group by Attributes

This defines how matching events are aggregated, only events with the same matching attribute values are grouped into one unique incident ID

jobId,jobName,jobAlgo,hostName,user,destName,jobDetail

Aggregate Constraint

This is most typically a numerical constraint that defines when the rule should trigger an incident

COUNT (*) >= 1

Incident Attribute Mapping

This section defines which fields in matching raw events should be mapped to the incident attributes in the resulting incident.

The available raw event attributes to map are limited to the group by attributes and the aggregate event constraint fields for each subpattern

 jobId = Anomaly.jobId,
 jobName = Anomaly.jobName,
 jobAlgo = Anomaly.jobAlgo,
 hostName = Anomaly.hostName,
 user = Anomaly.user,
 destName = Anomaly.destName,
 jobDetail = Anomaly.jobDetail

FortiSIEM Rules