Manage your SOC better with incident metrics

Important

The incident metrics features are currently in public preview. These features are provided without a service level agreement, and are not recommended for production workloads. For more information, see Supplemental Terms of Use for Microsoft Azure Previews.

As a Security Operations Center (SOC) manager, you need to have overall efficiency metrics and measures at your fingertips to gauge the performance of your team. You'll want to see incident operations over time by many different criteria, like severity, MITRE tactics, mean time to triage, mean time to resolve, and more. Azure Sentinel now makes this data available to you with the new SecurityIncident table and schema in Log Analytics and the accompanying Security operations efficiency workbook. You'll be able to visualize your team's performance over time and use this insight to improve efficiency. You can also write and use your own KQL queries against the incident table to create customized workbooks that fit your specific auditing needs and KPIs.

Use the security incidents table

The SecurityIncident table is built into Azure Sentinel. You'll find it with the other tables in the SecurityInsights collection under Logs. You can query it like any other table in Log Analytics.

Security incidents table

Every time you create or update an incident, a new log entry will be added to the table. This allows you to track the changes made to incidents, and allows for even more powerful SOC metrics, but you need to be mindful of this when constructing queries for this table as you may need to remove duplicate entries for an incident (dependent on the exact query you are running).

For example, if you wanted to return a list of all incidents sorted by their incident number but only wanted to return the most recent log per incident, you could do this using the KQL summarize operator with the arg_max() aggregation function:

SecurityIncident
| summarize arg_max(LastModifiedTime, *) by IncidentNumber

More sample queries

Mean time to closure:

SecurityIncident
| summarize arg_max(TimeGenerated,*) by IncidentNumber 
| extend TimeToClosure =  (ClosedTime - CreatedTime)/1h
| summarize 5th_Percentile=percentile(TimeToClosure, 5),50th_Percentile=percentile(TimeToClosure, 50), 
  90th_Percentile=percentile(TimeToClosure, 90),99th_Percentile=percentile(TimeToClosure, 99)

Mean time to triage:

SecurityIncident
| summarize arg_max(TimeGenerated,*) by IncidentNumber 
| extend TimeToTriage =  (FirstModifiedTime - CreatedTime)/1h
| summarize 5th_Percentile=max_of(percentile(TimeToTriage, 5),0),50th_Percentile=percentile(TimeToTriage, 50), 
  90th_Percentile=percentile(TimeToTriage, 90),99th_Percentile=percentile(TimeToTriage, 99) 

Security operations efficiency workbook

To complement the SecurityIncidents table, we’ve provided you an out-of-the-box security operations efficiency workbook template that you can use to monitor your SOC operations. The workbook contains the following metrics:

  • Incident created over time
  • Incidents created by closing classification, severity, owner, and status
  • Mean time to triage
  • Mean time to closure
  • Incidents created by severity, owner, status, product, and tactics over time
  • Time to triage percentiles
  • Time to closure percentiles
  • Mean time to triage per owner
  • Recent activities
  • Recent closing classifications

You can find this new workbook template by choosing Workbooks from the Azure Sentinel navigation menu and selecting the Templates tab. Choose Security operations efficiency from the gallery and click one of the View saved workbook and View template buttons.

Security incidents workbook gallery

Security incidents workbook complete

You can use the template to create your own custom workbooks tailored to your specific needs.

SecurityIncidents schema

The data model of the schema

Field Data type Description
AdditionalData dynamic Alerts count, bookmarks count, comments count, alert products names and tactics
AlertIds dynamic Alerts from which incident was created
BookmarkIds dynamic Bookmarked entities
Classification string Incident closing classification
ClassificationComment string Incident closing classification comment
ClassificationReason string Incident closing classification reason
ClosedTime datetime Timestamp (UTC) of when the incident was last closed
Comments dynamic Incident comments
CreatedTime datetime Timestamp (UTC) of when the incident was created
Description string Incident description
FirstActivityTime datetime First event time
FirstModifiedTime datetime Timestamp (UTC) of when the incident was first modified
IncidentName string Internal GUID
IncidentNumber int
IncidentUrl string Link to incident
Labels dynamic Tags
LastActivityTime datetime Last event time
LastModifiedTime datetime Timestamp (UTC) of when the incident was last modified
(the modification described by the current record)
ModifiedBy string User or system that modified the incident
Owner dynamic
RelatedAnalyticRuleIds dynamic Rules from which the incident's alerts were triggered
Severity string Severity of the incident (High/Medium/Low/Informational)
SourceSystem string Constant ('Azure')
Status string
TenantId string
TimeGenerated datetime Timestamp (UTC) of when the current record was created
(upon modification of the incident)
Title string
Type string Constant ('SecurityIncident')

Next steps