Schedule Trigger¶

Trigger

The Schedule Trigger is fired by the Pipelit scheduler system at configured intervals. It enables recurring workflow execution without external cron -- interval-based scheduling with retry logic, pause/resume, and automatic crash recovery.

Component type: trigger_schedule

Ports¶

Outputs¶

Inputs¶

This component has no input ports. It is an entry point.

Configuration¶

The Schedule Trigger itself requires no canvas-level configuration. Scheduling is managed through ScheduledJob records, which are created and controlled via the Schedules API.

ScheduledJob Properties¶

Trigger-Level Matching¶

The trigger's configuration supports an optional scheduled_job_id field. When set, the trigger only fires for events from that specific ScheduledJob. This allows multiple Schedule Triggers on the same workflow to be targeted by different schedules.

Usage¶

1. Drag a Schedule Trigger onto the canvas.
2. Connect it to downstream nodes that should execute on a recurring basis.
3. Create a ScheduledJob via the API that targets this workflow and trigger node.

Creating a Scheduled Job¶

POST /api/v1/schedules/

{
  "workflow_id": 1,
  "trigger_node_id": "trigger_schedule_abc123",
  "interval_seconds": 3600,
  "repeat_count": 0,
  "retry_max": 3
}

This creates a job that fires every hour, repeats indefinitely, and retries up to 3 times on failure.

Managing Schedules¶

Accessing Schedule Data Downstream¶

Reference the trigger timestamp in downstream nodes:

{{ trigger.timestamp }}

The payload also contains the scheduled_job_id for identifying which schedule triggered the execution.

How the Scheduler Works¶

The Pipelit scheduler uses a self-rescheduling architecture built on RQ (Redis Queue):

1. A ScheduledJob is created and enqueued as an RQ job.
2. When the job fires, it dispatches the workflow via dispatch_event("schedule", ...).
3. After execution, the scheduler calls _enqueue_next() to schedule itself again using Queue.enqueue_in().
4. Each enqueued job gets a deterministic RQ job ID (sched-{id}-n{repeat}-rc{retry}) to prevent duplicate enqueues.
5. On failure, exponential backoff is applied, capped at 10x the interval.
6. On startup, recover_scheduled_jobs() re-enqueues any active jobs whose next_run_at is in the past.

Example¶

A scheduled monitoring workflow:

graph LR
    ST[Schedule Trigger] --> AG[Agent]
    AM[AI Model] -.-> AG
    SH[System Health] -.-> AG
    HR[HTTP Request] -.-> AG

    style ST fill:#f97316,color:white
    style AM fill:#3b82f6,color:white
    style SH fill:#10b981,color:white
    style HR fill:#10b981,color:white

1. Schedule Trigger fires every hour.
2. Agent checks system health and makes HTTP requests to external services.
3. Results are logged or sent via a delivery channel.

Scheduled report with Telegram delivery¶

graph LR
    ST[Schedule Trigger] --> CD[Code]
    CD --> AG[Agent]
    AM[AI Model] -.-> AG

    style ST fill:#f97316,color:white
    style AM fill:#3b82f6,color:white

A daily report workflow: the Schedule Trigger fires at a set interval, a Code node gathers data, and an Agent summarizes and formats the report for delivery.