detection.wiki

Detection rules › Kusto

Dataverse - New Dataverse application user activity type

Status
available
Severity
medium
Time window
14d
Group by
EntityName, InstanceUrl, Message, UserId
Source
github.com/Azure/Azure-Sentinel

Identifies new or previously unseen activity types associated with Dataverse application (non-interactive) user.

MITRE ATT&CK coverage

TacticTechniques
Privilege Escalation
Credential Access
Execution

Telemetry coverage

PlatformRecord / event type
Microsoft 365any: Dataverse and model-driven apps audit (catch-all)

Rule body

id: 5c768e7d-7e5e-4d57-80d4-3f50c96fbf70
kind: Scheduled
name: Dataverse - New Dataverse application user activity type
description: Identifies new or previously unseen activity types associated with Dataverse
  application (non-interactive) user.
severity: Medium
status: Available
requiredDataConnectors:
  - connectorId: Dataverse
    dataTypes:
      - DataverseActivity
queryFrequency: 1h
queryPeriod: 14d
triggerOperator: gt
triggerThreshold: 0
tactics:
  - CredentialAccess
  - Execution
  - PrivilegeEscalation
relevantTechniques:
  - T1635
  - T0871
  - T1078
query: |
  let query_frequency = 1h;
  let query_lookback = 14d;
  let app_user_regex = "^[0-9A-Fa-f]{8}-[0-9A-Fa-f]{4}-[0-9A-Fa-f]{4}-[0-9A-Fa-f]{4}-[0-9A-Fa-f]{12}\\.com$";
  let guid_regex = "([0-9A-Fa-f]{8}-[0-9A-Fa-f]{4}-[0-9A-Fa-f]{4}-[0-9A-Fa-f]{4}-[0-9A-Fa-f]{12})";
  let application_users = DataverseActivity
      | where UserId !endswith "@onmicrosoft.com" and UserId != "Unknown"
      | summarize by UserId
      | where split(UserId, "@")[1] matches regex app_user_regex;
  let historical_app_activity = application_users
      | join kind = inner (
          DataverseActivity
          | where TimeGenerated between(ago(query_lookback) .. ago(query_frequency))
          | summarize by UserId, EntityName, Message, InstanceUrl)
          on
          UserId;
  let current_activity = application_users
      | join kind= inner (
          DataverseActivity
          | where TimeGenerated >= ago(query_frequency)
          | summarize by UserId, EntityName, Message, InstanceUrl)
          on
          UserId;
  current_activity
  | join kind = leftanti (historical_app_activity) on UserId, Message, EntityName, InstanceUrl
  | summarize NewActivities = make_set(strcat(Message, " ", EntityName), 1000) by UserId, InstanceUrl
  | extend
      AadUserId = extract(guid_regex, 1, tostring(split(UserId, "@")[0])),
      CloudAppId = int(32780)
  | project
      UserId,
      NewActivities,
      InstanceUrl,
      AadUserId,
      CloudAppId
eventGroupingSettings:
  aggregationKind: SingleAlert
entityMappings:
  - entityType: Account
    fieldMappings:
      - identifier: AadUserId
        columnName: AadUserId
  - entityType: CloudApplication
    fieldMappings:
      - identifier: AppId
        columnName: CloudAppId
      - identifier: InstanceName
        columnName: InstanceUrl
alertDetailsOverride:
  alertDisplayNameFormat: 'Dataverse - Unusual non-interactive account activity in
    {{InstanceUrl}} '
  alertDescriptionFormat: '{{UserId}} generated new activities in {{InstanceUrl}}
    which had not been seen previously in the Dataverse.'
version: 3.2.0

Stages and Predicates

Parameters

let query_frequency = 1h;
let query_lookback = 14d;
let app_user_regex = "^[0-9A-Fa-f]{8}-[0-9A-Fa-f]{4}-[0-9A-Fa-f]{4}-[0-9A-Fa-f]{4}-[0-9A-Fa-f]{12}\\.com$";
let guid_regex = "([0-9A-Fa-f]{8}-[0-9A-Fa-f]{4}-[0-9A-Fa-f]{4}-[0-9A-Fa-f]{4}-[0-9A-Fa-f]{12})";

let application_users and let current_activity are inlined into the numbered stages below.

Let binding: historical_app_activity used in Stage 6

let historical_app_activity = application_users
    | join kind = inner (
        DataverseActivity
        | where TimeGenerated between(ago(query_lookback) .. ago(query_frequency))
        | summarize by UserId, EntityName, Message, InstanceUrl)
        on
        UserId;

Stages 1 to 5 define let current_activity (the rule's main pipeline source); stages 6 to 9 run on it.

Stage 1: source

DataverseActivity

Stage 2: where

| where UserId !endswith "@onmicrosoft.com" and UserId != "Unknown"

Stage 3: summarize

| summarize by UserId

Stage 4: where

| where split(UserId, "@")[1] matches regex app_user_regex

Stage 5: join

| join kind= inner (
        DataverseActivity
        | where TimeGenerated >= ago(query_frequency)
        | summarize by UserId, EntityName, Message, InstanceUrl)
        on
        UserId

Stage 6: join (negated)

current_activity
| join kind = leftanti (historical_app_activity) on UserId, Message, EntityName, InstanceUrl

Stage 7: summarize

| summarize NewActivities = make_set(strcat(Message, " ", EntityName), 1000) by UserId, InstanceUrl

Stage 8: extend

| extend
    AadUserId = extract(guid_regex, 1, tostring(split(UserId, "@")[0])),
    CloudAppId = int(32780)

Stage 9: project

| project
    UserId,
    NewActivities,
    InstanceUrl,
    AadUserId,
    CloudAppId

Exclusions

The rule actively suppresses these predicates.

FieldKindExcluded valuesSearch
UserIdends_with@onmicrosoft.comexcludes:UserId field:"UserId" value:"@onmicrosoft.com"
UserIdneUnknownexcludes:UserId field:"UserId" value:"Unknown"

Indicators

These rows show field, operator, and value matches.

FieldKindValuesSearch
UserIdends_with
  • @onmicrosoft.com
field:"user" kind:ends_with value:"@onmicrosoft.com"
UserIdne
  • Unknown
field:"user" kind:ne value:"Unknown"

Output fields

These fields are emitted when the rule matches.

FieldSource
AadUserIdproject
CloudAppIdproject
InstanceUrlproject
NewActivitiesproject
UserIdproject