Scheduled jobs

Scheduled runs work on a cron, an interval or at one exact moment, inside the Tusk process — no external cron, no sidecar. Jobs, their runs and their output live in ~/.tusk/.

Job kinds

Kind	What it does
Backup	pg_dump of a connection to a destination folder, in custom, plain or directory format. Keep last N prunes older files, so a daily backup does not fill the disk.
Vacuum / Analyze	Maintenance on a whole database or a list of tables.
Query	Run a SQL statement on a connection. Results can be saved as a snapshot (Save results as) and downloaded from the run history — a cheap way to keep a daily report.
Pipeline	Execute a pipeline saved on the Data page; each run's output file is kept and downloadable.
Schema watch	Snapshot the catalog and notify on drift — see Schema Watch.
Plugin	A job a plugin registered (Analytics uses this for dashboard refreshes).

Triggers

• Cron — a standard five-field expression (0 6 * * *). The page shows the next run in your timezone.
• Interval — every N seconds, minutes or hours.
• One-time — at an exact moment; the job is removed after it runs.

A job can be paused and resumed, run on demand (Run now) and its trigger changed without recreating it.

Runs and notifications

Every run is recorded with start, duration, status and the error message when it failed. Each run ends with a scheduler.job.success or scheduler.job.error event; subscribe a channel to the error one under Notifications to get Slack, Discord, Telegram, email or a webhook call. Backups additionally emit core.backup.completed / core.backup.failed.

API

GET    /api/scheduler/jobs
POST   /api/scheduler/jobs/{backup|vacuum|analyze|query|pipeline|schema_watch|plugin}
POST   /api/scheduler/jobs/{id}/run | pause | resume
PUT    /api/scheduler/jobs/{id}/trigger
GET    /api/scheduler/jobs/{id}/runs
DELETE /api/scheduler/jobs/{id}

Jobs are owned by the user who created them in multi-user mode; admins see all of them.