Sign in
jobs

Get Job per-file processing status

Changed on

Retrieve the processing status of every file in a job: one entry per file, carrying its status for the whole job and the workflow step it currently sits at. Complements /jobs/{job_id}/details, which reports the same work as per-node counts, and /jobs/{job_id}/failed-files, which covers failures only. Call this endpoint at most 12 times per minute for each job and API key or JWT identity. For source-listed files, filter by discovered_by_node_id; that subset is complete when index_finished is true and truncated is false. Read the job status alongside this response because a failed source also counts as finished.

get/api/v1/jobs/{job_id}/file-status

Request

  • Base URL: https://platform.unstructuredapp.io
  • URL: https://platform.unstructuredapp.io/api/v1/jobs/{job_id}/file-status
  • Auth: none declared

Path parameters

job_idstring uuid required

Headers

unstructured-api-keystring nullable

Response

Successful Response

idstring uuid required
truncatedboolean

True when the data plane hit its per-job file cap, so this is a subset of the job's files rather than all of them. The cap selects files in id order, so the subset is arbitrary -- treat an absent file as unknown, not as missing.

index_finishedboolean nullable

True once no source connector is still listing files, so no further files will arrive from one. A source whose node failed counts as finished. A listing that failed while its node is still live does not: it is retried, so a failed job can end with this false. Read the job's own status alongside it. Not a promise that this list is final: a workflow step can still introduce files afterwards, which discovered_by_node_id (on each file) tells apart from what a source listed. Null when the job cannot say, which today means a resident job, or when the data plane does not report it.

generated_atstring date-time nullable

When the data plane produced the underlying report; it trails live execution.

Changes