adayush commented on issue #38216:
URL: https://github.com/apache/superset/issues/38216#issuecomment-3958225256
This is all I've tried:
### What we observed
* One report schedule **never triggers** (no Celery logs, no email, no UI
log).
* Another report in the **same environment works reliably**.
* Celery Beat and workers are up and healthy.
---
### Environment
* **Apache Superset v5**
* Celery + Redis
* Celery Beat running in a **separate container**
* Reports scheduled with timezone `Asia/Calcutta`
---
### What we tried (chronological)
1. **Verified Celery Beat health**
* Beat logs show `Scheduler: Sending due task reports.scheduler` every
minute.
* Confirmed this is expected behavior (global scheduler tick).
2. **Found and fixed Celery Beat PID issue**
* Beat was failing due to stale `/tmp/celerybeat.pid`.
* Fixed by restarting container (`docker compose down beat && up -d`).
* After this, Beat logs and scheduler task execution resumed normally.
3. **Confirmed scheduler task execution**
* Worker logs show:
* `reports.scheduler received`
* `reports.scheduler succeeded`
* For the broken report, **no `reports.execute` task is ever enqueued**.
4. **Compared working vs non-working reports**
* Working report:
* Has `last_eval_dttm` populated
* `last_state = Success`
* Generates `report_schedule_log` entries
* Broken report:
* `last_eval_dttm = NULL`
* `last_state = Not triggered` (initially)
* No rows ever appear in `report_schedule_log`
5. **Reset execution state of the broken report**
* Manually set:
* `last_state = NULL`
* `last_eval_dttm = NULL`
* `last_value = NULL`
* `last_value_row_json = NULL`
* Result: **no change**, still no Celery execution.
6. **Changed cron schedules**
* Updated cron times via UI (e.g. 3:25 → 4:59).
* Also tested future times and minute-level schedules.
* Result: **still no Celery logs** for this report.
7. **Verified time & timezone**
* Confirmed `date` is identical in:
* Superset web container
* Celery worker
* Celery beat
* Timezone (`Asia/Calcutta`) has been working historically.
8. **Validated dashboard / chart**
* Confirmed the chart and dashboard exist.
* Re-selected the **same dashboard as the working report**.
* Saved and retried scheduling.
* Result: **still skipped silently**.
9. **Verified recipients**
* Confirmed email recipients exist for the broken report.
* Compared recipient tables with the working report — no issue found.
10. **Observed final DB state (broken report)**
* `active = true`
* `crontab` correct
* `timezone = Asia/Calcutta`
* `dashboard_id` valid
* `last_eval_dttm = NULL`
* `last_state = NULL`
* No entries in `report_schedule_log`
* No `reports.execute` task ever logged by Celery
---
### Key symptom to highlight
* `reports.scheduler` **runs successfully**
* Other reports in the same system work.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
