Scheduler
The Scheduler runs any notebook on a cron expression — no separate orchestrator to deploy.
Creating a scheduled job
Section titled “Creating a scheduled job”- Open Scheduler and create a new job.
- Pick the notebook to run and a cron expression for when it should fire.
- Optionally set a
max_run_timeout_secondsso a stuck run doesn’t block the next one.
Dependent jobs
Section titled “Dependent jobs”A job can be configured to run only after another job completes successfully — chain ingestion → transform → dashboard-refresh jobs without a separate DAG tool.
Run history
Section titled “Run history”Every fire is recorded with the full cell code and output from that run, so you can debug a failure without re-running the notebook manually. A live countdown shows time until the next scheduled fire.
Triggering outside the cron schedule
Section titled “Triggering outside the cron schedule”Every job also gets a webhook endpoint (POST /webhook/{token}) so external systems can trigger a run on demand, alongside the authenticated POST /{job_id}/run endpoint.
Concurrency
Section titled “Concurrency”Org-level max_concurrent_jobs caps how many scheduled runs execute at once, so a burst of due jobs can’t overwhelm shared runtime capacity.
Next steps
Section titled “Next steps”- Dashboards to visualize what a scheduled job produced
- ArrowLake for a good target table for scheduled writes