Documentation

A narrowing predicate on a job listing. FilterJobId and the completed_at range name DLQ and archive columns.

Sortable columns on the main jobs table.

Underlying SQL column name for a JobSortColumn.

Sortable columns on the DLQ table. DlqId is the DLQ primary key and DlqJobId the original job id.

The DLQ column a sort key names.

Sortable columns on the archive table. ArchiveId is the archive primary key and ArchiveJobId the original job id.

The archive column a sort key names.

Sort direction.

A sort direction as SQL.

A throttle-deferred job with a live marker, still parked. Shared by status, count, and wake. Claiming clears the marker.

The derived job status. Its string values match jobStatusToText.

All job columns plus the derived status column, aliased job, for filtering.

List filtered jobs without the derived status.

List filtered jobs with status as a trailing column. The caller attaches the row decoder.

Count filtered jobs through the status subquery.

Fetch a single job by id with its derived status trailing column. The caller attaches the row decoder.

List DLQ jobs under a dynamic WHERE and an orderBy from buildDLQOrderBy.

How NULLs in a sort column should order relative to non-NULL values.

The NULLS placement for a column's null semantics and direction.

How nulls sort for a job column.

How nulls sort for a DLQ column.

ORDER BY for the jobs table. Defaults to id DESC.

ORDER BY for the DLQ table. Defaults to failed_at DESC.

ORDER BY for the archive table. Defaults to completed_at DESC.

Count DLQ jobs under a dynamic WHERE.

The DLQ read columns, in codec order. The DLQ uses job_id for the main-table id.

The job read columns except id. The archive INSERT copies them with the main table's id as job_id.

Job columns carried through a DLQ round-trip. The read columns except id and last_error, plus write-only rate_limit_cost.

Job columns a DLQ retry carries back to the main table. The retry re-arms the rest.

requeuedCols for an archive re-enqueue, without the parent link.

The job read columns, in codec order, for SELECT and RETURNING.

DO UPDATE SET body for a replace-dedup upsert. Copies each writable column from the excluded row, then re-arms the replaced job for a fresh run.

Insert a job. The write fragment carries the column list and parameters.

Insert under the replace dedup strategy. Replaces an existing job that is idle and childless in the main queue and the DLQ. ON CONFLICT DO UPDATE fires the groups UPDATE trigger, which maintains a cross-group move.

Batch insert over unnested parallel arrays. An ignore-dedup job is skipped on conflict. A replace-dedup job updates an idle existing row.

insertJobsBatchBase with no RETURNING.

Batch insert with dedup and replaceable-job handling, plus a caller's RETURNING.

Fetch a job by id.

Fetch a job by its dedup key. The partial unique index guarantees at most one row.

Delete a job by id. Refuses one with children, which cancelJobCascadeSQL takes. A deleted child with no siblings left resumes its parent for a completion round.

UNION ALL of body over each job table, passing its raw name and schema-qualified reference.