DataHub identified a defect affecting scheduled ingestion runs on versions v0.3.16.x and v0.3.17.x. In some cases, scheduled runs were being misclassified as CLI (command line) runs, causing them to disappear from the ingestion run history UI — even though ingestion was completing successfully in the background.
If you use DataHub assertions, you may have also seen assertion failures during this period as a result.
If you self-host a Remote Executor, resolving this requires two steps:
- Upgrade GMS (Generalized Metadata Service) to the hotfix version for your release line:
- On v0.3.16.x → upgrade to v0.3.16.7
- On v0.3.17.x → upgrade to v0.3.17.2
- Upgrade your Remote Executor to images matching your respective GMS version and restart the executor process.
Both steps are required — the GMS (Generalized Metadata Service) upgrade alone will not fully resolve the issue on your executor.
After both are applied, the next scheduled run should appear correctly in the run history UI, and any affected assertions should recover automatically.
If you have additional questions about how to resolve this issue, please feel free to log a ticket with DataHub Support here.
Additional Reference: DataHub Incident #2715