Jobs

Start, follow and cancel large runs

Every background job is visible, resumable and controllable — including after a restart.

curl "$API/jobs/lanes"

{
  "lanes": [
    { "lane": "statements.esef", "running": 1, "backlog": 412 },
    { "lane": "earnings.dates.repair", "running": 0, "backlog": 0 }
  ]
}

What you can do with it

Watch or cancel a running jobCancellation takes effect within seconds.
See the backlog per laneOne call, all lanes.
Trigger everything that is duePOST /sync/all — never blocks.

Why we built it this way

A job whose process died used to sit on “running” forever — blocking every further run of the same lane.

Every job now carries the identity of the process that started it. Whoever does not hold that identity cannot still be running and is recorded accordingly. That is the whole idea, and it cleared twelve stale records on the first start.

Runs are derived from the data itself, not from an offset. A cancelled run therefore resumes where it stood, with no bookkeeping.

Improvements

Crashed jobs are detected automatically and recorded correctly
Jobs can be cancelled
Runs resume after a cancellation instead of starting over
New single call for everything that is due

Endpoints

GET	`/jobs/lanes`	Backlog per lane
GET	`/jobs/{job_id}`	Status of a job
POST	`/jobs/{job_id}/cancel`	Cancel a job
POST	`/sync/all`	Everything that is due

Give feedback or ask a question

Something missing, something off, or an answer you don't understand? Write it here — it reaches us directly.