API ReferenceObservability

Start a DAG-failure diagnosis run

Starts an asynchronous AI-powered diagnosis run for a failed DAG run (or specific task within it) and returns a run ID. Use the run ID to poll status and stream events.

Authentication

AuthorizationBearer

Bearer authentication of the form Bearer <token>, where token is your auth token.

Path parameters

organizationIdstringRequired
The ID of the Organization to which the Deployment belongs.
deploymentIdstringRequired
The Deployment's ID.

Query parameters

forcebooleanOptionalDefaults to false

Skip DB session dedup + Redis cache and force a fresh diagnosis run

Request

The request body for starting a DAG-failure diagnosis run.

dagIdstringRequired
The ID of the DAG that failed.
runIdstringRequired
The ID of the DAG run that failed.
skipCacheReadbooleanOptional

If true, bypass any previously cached diagnosis for this DAG/run/task and force a fresh run.

taskIdstringOptional

The ID of the task that failed. Optional; if omitted, the agent diagnoses the run as a whole.

tryNumberslist of integersOptional
The specific task try numbers to diagnose. Optional.

Response

OK
createdAtstring
The RFC3339 timestamp when the run was created.
runIdstring
The unique identifier of the diagnosis run. Use this to poll status and stream events.
statusobject
The current status of the diagnosis run.

Errors

400
Bad Request Error
401
Unauthorized Error
403
Forbidden Error
429
Too Many Requests Error
500
Internal Server Error