AuditLogRowDto
Example
{ "id": "string", "account_id": "string", "application_id": "string", "environment_id": "string", "actor_id": "string", "actor_type": "string", "action": "string", "resource_type": "string", "resource_id": "string", "metadata": {}, "created_at": "2026-04-20T12:00:00.000Z", "actor_label": "string", "resource_label": "string", "correlation_id": "00000000-0000-0000-0000-000000000000", "outcome": "success", "category": "auth", "severity": "info", "customer_visible": false, "identity_visible": false }
Properties
Snapshot label of the actor at write time (email for users / identities, label otherwise). NULL on rows written before the Slice 2 deploy.
Snapshot label of the affected resource at write time. NULL on rows written before the Slice 2 deploy.
Per-request correlation ID — every row written during the same HTTP request (or background job) shares this value. Use the query API's `correlation_id` filter to reconstruct a chain.
`success` for normal-flow events, `failure` for errors, `denied` for authorization rejections.
Broad classification from the action catalog (e.g. `auth`, `rbac`). NULL on legacy rows whose action key was unknown at backfill time — those carry the `unknown` value in storage.
`info` for normal-flow events, `notice` for important state changes, `warning` for failures, `critical` for security events.
Whether this row should appear in customer-facing audit views. `false` rows are internal-only (platform support).
Tell us how we can improve this guide.