Run workflow

Learn how to save and run the pipeline you design, as well as monitoring features.

Save

Be sure to save your changes before running. Click the Save button on the top toolbar or use the shortcut key Ctrl+S. If you entered the editor directly without going through the collection/folder, select the location first when saving.

validation

Connection and essential settings are checked before execution. If there is a problem, execution stops and displays the corresponding node.

Manual Run

batch Pipeline

1. Click the Run button (play icon) on the top toolbar.
2. When execution begins, the button changes to Stop, and the connection lines on the canvas move to show data flowing.
3. Click the Stop button to stop execution.

event Pipeline

1. Click the Register button on the top toolbar to register event reception. The event topic to subscribe to must be determined in advance in the pipeline settings.
2. The pipeline changes to Listening state and runs automatically when an event comes in.
3. To stop receiving events, click the Unregister button.

The Register/Unregister buttons only appear in the event pipeline, and the Run/Stop buttons appear in the batch pipeline.

Scheduling

To run a pipeline periodically:

1. Click Settings (gear icon) on the top toolbar.
2. Set the cycle in the Schedule tab.
3. Once you save your settings, you will see a Schedule badge in your pipeline list.

For detailed scheduling settings, refer to the Scheduling document.

Monitoring

Real-time status

During execution, status indication (icon, label, number of records processed, time required) appears above each node, and a soft light (glow) effect appears on the currently running node.

status	Description
running	Currently running
Success (ok / ready)	Successfully completed
failure (error / failed)	execution failed
pending	Wait for dependency node completion

Execution history bar

The Execution History Bar in the top toolbar shows recent runs of the saved pipeline in color.

• Each bar is an individual action, and success/failure is distinguished by color.
• When you hover your mouse over the bar, the execution number, status, time, and time required are displayed.
• Check batch-specific details and step logs in the History tab of the code node.

Deployment Details

If you select a deployment in the History tab of a code node, you can see the following:

• Batch status and time required
• Status by step: Execution result and start/end time of each step
• Error Details: Description and stack trace of the failed step

Error handling

Execution failure notification

If execution fails, Error Notification (Toast) will appear on the screen and the reason for failure will be displayed. If you allow browser notifications, you will receive desktop notifications of completion or failure even when you are not looking at the editor (✓·✗ will also briefly appear in the browser tab title). Both Batch·Event pipelines operate the same.

• Errors for each step are checked through stack traces in the code node History tab (see below).
• On the other hand, for infrastructure failures that leave no step-by-step records, such as a running pod not being able to pop up, the run-level reason is displayed directly in this notification.

Execution with unknown state

If the execution status is not read in succession (approximately 1 minute), the editor will not repeat the query indefinitely, but will stop checking the status and announce Pipeline status unknown. In this case, check the network or backend status and run again.

Things to check when execution fails

1. Select the failing code node to open its properties panel.
2. Select the failed batch in the History tab.
3. Check the description and stack trace (record of error location) of the failed step (copyable).

Common error types

Error Type	Cause	Solution
ConnectionError	Data source connection failure	Network, check authentication information
Schema Mismatch	schema mismatch	Check input/output schema mapping
Code Error	Python/SQL code error	Code grammar, logic modification
Resource Limit	lack of resources	Reduce data size, adjust partition
Timeout	execution timeout	Query optimization, adjust timeout settings