apache/superset · warning · ReportScheduleUnexpectedError

Report schedule was modified or deleted by another process d

Error message

Report schedule was modified or deleted by another process during execution

What it means

ReportScheduleUnexpectedError wrapping a SQLAlchemy StaleDataError: while the execution was finishing, the UPDATE that persists the final log state failed because the report_schedule row's version changed or was deleted by another transaction. The code rolls back, logs a warning explaining the report was likely deleted mid-run, and re-raises as this error so the worker marks the execution failed cleanly.

Source

Thrown at superset/commands/report/execute.py:468

                    uuid=self._execution_id,
                )
                db.session.add(log)
            log.end_dttm = datetime.now(timezone.utc).replace(tzinfo=None)
            log.value = self._report_schedule.last_value
            log.value_row_json = self._report_schedule.last_value_row_json
            log.state = effective_state
            log.error_message = log_message
            db.session.commit()  # pylint: disable=consider-using-transaction
        except StaleDataError as ex:
            # Report schedule was modified or deleted by another process
            db.session.rollback()  # pylint: disable=consider-using-transaction
            logger.warning(
                "Report schedule (execution %s) was modified or deleted "
                "during execution. This can occur when a report is deleted "
                "while running.",
                self._execution_id,
            )
            raise ReportScheduleUnexpectedError(
                "Report schedule was modified or deleted by another process "
                "during execution"
            ) from ex

    def _get_url(
        self,
        user_friendly: bool = False,
        result_format: Optional[ChartDataResultFormat] = None,
        **kwargs: Any,
    ) -> str:
        """
        Get the url for this report schedule: chart or dashboard
        """
        chart = self._report_schedule.chart
        dashboard = self._report_schedule.dashboard

        # Soft delete removed the FK-level guarantee that a report's target
        # chart exists: ``chart`` is a visibility-filtered relationship, so a

View on GitHub (pinned to f4587218dd)

Solutions

No data fix needed — the report was intentionally changed/deleted; verify the schedule still exists and let the next fire run normally
Prevent overlap by spacing crontabs (and see ALERT_REPORTS_MINIMAL_INTERVAL) so executions don't race on the same row
If it recurs without manual deletes, audit for duplicate Celery beat schedules firing the same task twice

Defensive patterns

Strategy: try-catch

Try / catch

try:
    _set_log_state()
except ReportScheduleUnexpectedError as ex:
    if 'modified or deleted' in str(ex):
        logger.warning('report changed mid-run; safe to ignore')
        return  # not a data-loss situation
    raise

Prevention

Space schedule fires so executions of one report never overlap
Prefer editing reports between fires, not during; check the worker's active runs first
Ensure only one Celery beat scheduler instance is running to avoid duplicate fires

When it happens

Trigger: A user deletes or edits the report schedule (or its owner/relations) while a Celery execution of that schedule is in flight; two overlapping executions race on the same row.

Common situations: Deleting a misbehaving report right as its scheduled run starts; overlapping schedules after worker restarts; long screenshot runs spanning an admin's cleanup of old reports.

Related errors

AI-assisted analysis of apache/superset@f4587218dd (2026-08-14). Data as JSON: /api/errors/e9637e65d3c9137e. Report an issue: GitHub.