Skip to main content
Question

How to create an alert from a SOAR workflow


scuba-steve
Forum|alt.badge.img+2

I have a playbook that needs to generate an alert in a specific environment if certain criteria are met after an investigation has been performed. 

It seems this may be possible in two ways:

  1. Create a UDM event in SecOps and have a detection trigger on it.
  2. Create an alert using the SOAR API.

It seems both would require a custom integration as these actions/endpoints are not available in the  default integrations available.

Which would be preferable? I feel as though going via a detection would be proper, as it would leave a full trail from event to alert. Using the SOAR API would probably be faster and easier however. 

I’ve been unable to find strong documentation on how to do either, hence the post, thanks!

2 replies

AymanC
Forum|alt.badge.img+14
  • Bronze 5
  • August 18, 2026

Hi ​@scuba-steve,.

 

The following PRE-GA offering may be something of interest:

 

Threat Hunt agent overview  |  Google Security Operations  |  Google Cloud Documentation

 

Otherwise, ‘User Requests’ is something quite nifty which may be something that is useful for you, an out of the box functionality to create a case, based on a pre-defined request. Once this request is raised, you can create a playbook to alert on that Request’s trigger value, so it automatically attaches a playbook to the case.

Create user requests  |  Google Security Operations  |  Google Cloud Documentation

 

Kind Regards,

Ayman


whathehack81
Forum|alt.badge.img+9

I’d lean toward the UDM → detection route if the desired output is genuinely a SecOps alert/detection rather than simply another SOAR work item.

The current Chronicle API exposes operations for existing CaseAlerts (get/list/move/update, etc.), but I don’t see a supported general-purpose create operation for CaseAlerts. That fits the normal SecOps model where alerts enter through ingestion/detection and are then grouped into cases.

In that design the playbook could emit a purpose-built UDM event containing the investigation result, environment, correlation identifiers, and relevant entities. A narrowly scoped YARA-L rule can then create the detection/alert. That also preserves the SIEM-side event → detection → alert trail you mentioned.

I’d avoid using User Requests for this particular workflow. Those are designed to create Request Cases from user-submitted request templates, rather than to generate an alert as the result of an investigation.

If you only need to surface the investigation outcome to analysts and don’t specifically require an alert/detection object, then updating the existing case, adding an insight, or creating another workflow artifact would be lighter than generating synthetic telemetry.