Get Job Status By Dataset

Look up the most recent ingest job for a dataset.

Used by the dataset detail page to surface ingest warnings permanently (S3 completion) — the job is the source of truth for reserved_rename / dbf_truncation_collision / mercator_clip / archive_failed / temporal_parse_errors metadata.

Returns the most recently created completed job for the dataset. When the dataset is visible but has no ingest job (e.g. registered from an existing table, or a remote/STAC dataset), returns 200 with a null body instead of 404 — a "no job" outcome is normal for these datasets and a 404 would needlessly pollute the browser console on the dataset detail page. A genuine 404 is still raised when the dataset is not visible to the user, to avoid leaking job existence (see visibility check below).

Not every caller gets every field. Seeing the dataset decides whether there is an answer at all; who ran the job decides how much of the answer is filled in. The dataset's owner, an admin, and the job's own creator get the full payload. Any other reader of a visible dataset gets the job id, its status and its timestamps, with the run's own detail nulled: no error_message, source_filename, warnings, step, row counts or retry hint. That is the redaction GET /datasets/{dataset_id}/refresh-runs already applies to the same failure text, and _redacted_job_status documents the decision field by field.