Schedule

The Schedule block automatically triggers workflows on a recurring schedule at specified intervals or times.

Use a Schedule trigger to run the same deployed Workflow on a cadence. Use a scheduled Sub Agent to repeat saved instructions against current workspace state. Sub Agent tasks have their own model, supported reasoning options, attached context, and run history. A workflow schedule uses the deployed blocks and each Agent block's model configuration.

Schedule Block

Schedule Options

Configure when your workflow runs:

  • Every X Minutes: Run at minute intervals (1-1440)
  • Hourly: Run at a specific minute each hour
  • Daily: Run at a specific time each day
  • Weekly: Run on a specific day and time each week
  • Monthly: Run on a specific day and time each month

Timezone

Daily, weekly, monthly, and cron schedules run in the Timezone you select. The dropdown offers 35 common zones and pre-selects your browser's timezone when you add the block. Every X Minutes and Hourly schedules do not use wall-clock time. SteelEngine hides the Timezone field for them.

Activation

Schedules are tied to workflow deployment:

  • Deploy workflow → Schedule becomes active and starts running
  • Undeploy workflow → Schedule is archived and stops running
  • Redeploy workflow → The schedule for each Schedule block is reused or recreated with the current configuration

You must deploy your workflow for the schedule to start running. Configure the schedule block, then deploy from the toolbar.

Execution Timeout

Each scheduled occurrence uses the workflow's Execution timeout, if set. The clock starts when SteelEngine admits that occurrence for execution, before its first block, rather than at the scheduled timestamp. A delayed occurrence therefore receives its full configured budget from admission. Waits, approvals, and subsequent worker delays count toward that deadline.

Changing or clearing the setting affects future occurrences; it does not extend one already running. Leave the setting blank to use the platform's async execution default.

Automatic Disabling

Schedules automatically disable after 100 consecutive failures to prevent runaway errors. When disabled:

  • A warning badge appears on the schedule block
  • The schedule stops executing
  • Click the badge to reactivate the schedule

Some failures disable the schedule immediately, without waiting for the counter:

  • An authentication or authorization error (HTTP 401 or 403), for example when the workflow is no longer deployed
  • A missing workflow (HTTP 404)

Fix the cause. Redeploy the workflow. Then reactivate the schedule.

Disabled Schedule

Schedule blocks cannot receive incoming connections and serve as workflow entry points only.

Common Questions

Yes. Schedules are created in the database only when you deploy the workflow. Undeploying archives the schedule so it stops firing, and redeploying reuses or recreates it with the current configuration.
After 100 consecutive failures, the schedule is automatically set to a disabled status to prevent runaway errors. A warning badge appears on the schedule block in the editor. You can click the badge to reactivate it. The failure counter resets to zero on any successful execution.
Yes. The schedule configuration includes a Timezone dropdown that pre-selects your browser's timezone when you add the block. Daily, weekly, monthly, and cron schedules are evaluated in that timezone; a schedule with no stored timezone falls back to UTC. Every X Minutes and Hourly schedules do not use a timezone.
If a rate limit (HTTP 429) is encountered during execution, the schedule automatically retries after a 5-minute delay rather than counting it as a failure.
Yes. The deployment process finds all schedule blocks in the workflow and creates a separate schedule record for each one. Each schedule operates independently with its own cron expression and failure counter.
The currently running execution will complete, but no new executions will be triggered. When the schedule next tries to fire, it checks that the workflow is still deployed and the schedule record still exists before executing.

On this page