Scheduling your first task

A task is a shell command that RunWisp starts on a schedule, or when you ask, and that exits when it’s done. Open the runwisp.toml from Installation and replace [tasks.hello] with your own job:

runwisp.toml

[tasks.backup-db]
cron       = "0 2 * * *"      # every day at 02:00
on_overlap = "skip"           # skip a run if the last one is still going
keep_runs  = 30               # keep the last 30 runs in the history
run        = "pg_dump mydb | gzip > /backups/mydb-$(date +%F).sql.gz"

cron is the usual five-field expression, in the host’s timezone. run goes to a shell, so pipes and $(...) work like they do in your terminal.

Apply the change

runwisp validate
runwisp reload

validate checks the file and starts nothing. reload gives the file to the running daemon, which starts the schedule. If the file has an error, the reload is rejected and the daemon keeps running what it had.

Run it now

You don’t have to wait for 02:00:

runwisp run backup-db

The output streams to your terminal, and the command exits with the task’s exit code. The run lands in the history like a scheduled one. In the TUI, the same thing is r on the task; in the Web UI, the Run now button.

Read its output

runwisp logs backup-db

Every run keeps its output, exit code, start time, and duration. In the Web UI, click the task to see its runs, and click a run to read its log. See Web UI and TUI.

Next steps

• Timezones, catching up after downtime, and spreading out start times are in Scheduling.
• What happens when a run is still going at the next start time is in Overlapping runs.
• Every key a task accepts is in the [tasks.*] reference.