Batch Scheduler Frequently Asked Questions (FAQ)
- In the application Batch Scheduler, why does a batch appear with the SUCCESS status in Batch Monitor even when Data Load or Scoring encounters errors?
Batch Monitor reflects SUCCESS or FAILURE statuses based on the run status returned to the batch framework by the underlying tasks. If a task encounters an error but handles it internally (for example, by recording it in a database table or writing it to logs) and does not propagate the error back to the framework, the framework may still mark the task (and therefore the batch) as successful.
- Can Batch Scheduler display a FAILURE status based on task execution outcomes?
Yes. The batch framework can show a SUCCESS or FAILURE status based on the run status of underlying tasks—provided the task reports the failure correctly, for example, by raising an exception that the framework can capture.
- What is the most common reason Data Load or Scoring fails in the back-end logs but the UI still displays a SUCCESS status?
The most common reason is that the Data Load or Scoring task encounters errors but does not report them to the batch framework as a task failure.
- How do we make Batch Monitor display a FAILED status when a Python or Notebook process contains errors?
Ensure that the task code raises an exception as soon as a fatal error is detected. When an exception is raised, the batch framework can capture it, mark the task as FAILED, and the Batch Monitor UI displays the status as FAILED.
- What about Notebook-based tasks?
For Notebook-based tasks, you must raise an exception in the appropriate notebook paragraph or cell where the failure is detected. The batch framework picks up the exception and marks the task as failed, and the UI displays the FAILED status.
- Is logging an issue in a database table sufficient to fail the batch?
Not by itself. Logging helps diagnostics, but the Batch Scheduler status depends on the task run status returned to the framework. If the code logs an error but does not propagate an exception or failure status, the batch may still be marked as successful.
- What must we verify when investigating a SUCCESS in UI but errors occurred report?
- Confirm whether the affected task, for example, Data Load or Scoring, logs exceptions but does not re-raise them.
- Review the execution status records or logs for the specific run to correlate what the code logged versus what status the framework recorded.
Logging Errors from Batch Execution
In OFS Compliance Studio’s Batch Scheduler, the Batch Monitor status is driven by the batch framework’s view of each task’s run outcome and not by what the task’s internal code records in logs. The framework determines success or failure based on whether each underlying task completes normally or returns a failure condition that the framework can recognize. As a result, it’s possible to see a batch with the SUCCESS status in the Batch Monitor UI even though the underlying Data Load or Scoring logic encountered errors in the back-end. This happens when the task code catches and handles errors internally, for example, recording an issue in a database table or writing the error to the logs, but does not propagate the error back to the batch framework as a task failure. In such a scenario, the batch framework has no failure signal to act on, so it continues to report the task and the overall batch, as successful.
To have the Batch Monitor accurately reflect a FAILED status when the underlying processing fails, the task implementation must explicitly signal the failure to the framework. For Notebook-based tasks, this typically means raising an exception in the relevant notebook paragraph or cell when a fatal condition is detected. When an exception is raised (instead of only being logged and then ignored), the batch framework can detect it, mark the task as failed, and the Batch Monitor UI will show the batch/task as FAILED along with the error details.