Jobs¶
The Jobs page displays workflow execution records and node states. It helps you determine when a workflow execution was triggered, its current or final state, and where a failure occurred.
Here, “jobs” covers only workflow execution records; it is not a unified list of all background tasks such as loads, exports, and SQL. View load tasks under Load Tasks.
Job list¶
The job list supports filtering by workflow name, execution state, and execution mode, and provides refresh, pagination, and details.
Column |
Description |
|---|---|
Workflow name |
The workflow to which this execution belongs |
Data |
The name of the data associated with this execution; empty if no data is associated |
State |
Preparing, running, paused, completed, failed, canceled, and other states |
Trigger method |
Manual, scheduled, or another execution mode |
Start time, end time |
The execution time range |
Processing duration |
Elapsed or total time spent on this execution |
Error message |
An error summary available when execution fails |
The job list refreshes records that are still running. If the displayed state differs from what you expect, use the page’s refresh action to retrieve the current execution state.
View job details¶
Click Details in the job list to view the full context of that workflow execution. Details primarily include basic information, the workflow topology, processed files, and node execution information.
Section |
Available information |
|---|---|
Basic information |
Workflow, execution ID, trigger method, state, start time, and end time |
Workflow topology |
Node execution states and their positions in the workflow |
Processed files |
Files produced or used by the current execution and the nodes they passed through |
Node details |
Node configuration, runtime input, output, variables, run summary, and error information |
Start with nodes that failed or took an unusually long time. Then use the node input, output, and error information to determine whether the issue lies in configuration, data, or an external dependency. To analyze a SQL statement, continue to SQL History.
Pause, resume, and rerun¶
The page displays actions according to the execution state and the operations returned by the backend. Not all jobs support the same set of operations; use only the operations currently shown in the UI.
Operation |
Use case |
|---|---|
Pause |
When the current execution can be paused, request a pause at an appropriate execution node |
Resume |
Continue a paused execution when resuming is supported |
Retry |
Start a new run for a failed execution and open the new execution details |
Rerun from node |
When supported on the details page, select a starting node and modify runtime parameters or continue according to the UI capabilities |
Read the failure information before rerunning. Retrying is appropriate for temporary resource shortages, network issues, or external-service fluctuations. Fix the root cause first for workflow configuration, input data, or permission issues.
State troubleshooting recommendations¶
Filter by the target workflow and execution state to find the abnormal execution record.
Open the details and confirm the failed node, error summary, and execution time.
Check whether the node’s input, output, variables, and configuration match expectations.
For SQL-related nodes, continue analyzing the query in SQL History.
For API submission or permission exceptions, view the current user’s request record in Operation Logs.