-
Notifications
You must be signed in to change notification settings - Fork 1.3k
Document agentic usage attribution in Audit Trail #37955
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: master
Are you sure you want to change the base?
Changes from all commits
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,60 @@ | ||
| --- | ||
| title: Track Agentic Usage in Your Organization | ||
| description: Use Audit Trail to see which actions Datadog AI agents take in your organization, distinguish agent activity from human activity, and trace agent actions back to the user who authorized them. | ||
| disable_toc: false | ||
| further_reading: | ||
| - link: "account_management/audit_trail/" | ||
| tag: "Documentation" | ||
| text: "Set up Audit Trail" | ||
| - link: "account_management/audit_trail/events/" | ||
| tag: "Documentation" | ||
| text: "Learn about Audit Trail events" | ||
| --- | ||
|
|
||
| ## Overview | ||
|
|
||
| Datadog AI agents, including the Bits fleet (Bits SRE, Bits Dev, Bits Assistant, Bits Security Analyst) and Agent Builder, can take actions inside your organization on their own or on behalf of a user. Audit Trail attributes these actions to the agent that performed them and, when applicable, to the user who authorized them. This guide walks you through finding agent-driven events, filtering them by agent type, and tracing an agent's actions back to the user who delegated them. | ||
|
|
||
| ## View all Datadog AI agent activity | ||
|
|
||
| 1. Navigate to [Audit Trail][1]. | ||
| 2. In the search bar, enter the query: `@evt.actor.mode:*`. | ||
|
|
||
| This query returns every event where a Datadog AI agent is the actor, whether the agent acted `interactive` (a user was present and approved the action) or `autonomous` (the agent acted on a system trigger, with no user present). | ||
| 3. Click an event to open details and see the full actor breakdown: the delegator (the user, if any), the agent, and the mode. | ||
|
|
||
| **Note**: Rollout to all Bits agents is in progress, and coverage is not yet complete. Actions from agents that have not yet onboarded do not appear in `@evt.actor.mode:*` results. | ||
|
|
||
| ## View actions taken through the Datadog MCP server | ||
|
|
||
| Actions taken through the Datadog MCP server, carry the `is_mcp_call` field. | ||
|
|
||
| 1. Navigate to [Audit Trail][1]. | ||
| 2. In the search bar, enter the query: `is_mcp_call:true`. | ||
| 3. Click an event to see a link back to the MCP server configuration and to other events from the same MCP session. | ||
|
|
||
| ## View a user's full activity, including agent-delegated actions | ||
|
|
||
| When a Datadog AI agent acts on a user's behalf, the agent becomes `@evt.actor` for that event, and the user moves to `@delegator.actor`. Searching only `@usr.name:<USER_EMAIL>` returns the actions the user performed directly, and does not include actions a Datadog AI agent performed on their behalf. | ||
|
Contributor
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. We might want to be a little more specific and mention that the agent now fills in the values of |
||
|
|
||
| To retrieve a user's complete activity, including delegated agent actions: | ||
|
|
||
| 1. Navigate to [Audit Trail][1]. | ||
| 2. In the search bar, enter the query: `(@usr.name:<USER_EMAIL> OR @delegator.name:<USER_EMAIL>)`. Replace `<USER_EMAIL>` with the user's email address. | ||
|
Contributor
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more.
|
||
|
|
||
| | Description of audit event | Query in audit explorer | | ||
| | ------------------------------------------------------------------------------------------------------| ------------------------------------------------------------ | | ||
| | Actions a user performed directly | `@usr.name:<USER_EMAIL>` | | ||
| | All actions attributed to a user, including actions a Datadog AI agent performed on their behalf | `(@usr.name:<USER_EMAIL> OR @delegator.name:<USER_EMAIL>)` | | ||
| 3. See [Create a dashboard or a graph][2] if you want to put this information into a dashboard or graph. | ||
|
|
||
| ## Distinguish agent events from human events | ||
|
|
||
| Agent-driven events display a visual badge in the event list, so you can distinguish them from human-driven events without opening each one. | ||
|
|
||
| ## Further reading | ||
|
|
||
| {{< partial name="whats-next/whats-next.html" >}} | ||
|
|
||
| [1]: https://app.datadoghq.com/audit-trail | ||
| [2]: /account_management/audit_trail/#create-a-dashboard-or-a-graph | ||
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.