Issue
Jobs that previously appeared in the Fusion UI under the Jobs tab are no longer visible. This may affect Spark, recommender, indexing, or signals aggregation jobs. However, when querying the Fusion API directly, these jobs still exist and return valid metadata.
Diagnosis
To verify if the jobs still exist despite not appearing in the UI, use the following API call:
curl -u <USER>:<PASSWORD> "https://<fusion-host>/api/apps/<app-name>/jobs"Replace <fusion-host> with your Fusion hostname and <app-name> with the name of your app. If jobs return successfully in the API response, but are not visible in the UI, the issue is likely due to a temporary disruption in the job-launcher microservice.
Environment
Fusion 5.x (Lucidworks-hosted and self-managed Kubernetes deployments)
Cause
This issue can occur when the job-launcher pod enters an inconsistent state or experiences a temporary failure. Although job metadata remains accessible via the API, the UI depends on the job-launcher service to populate the Jobs tab. A stale or stalled pod may result in partial or missing data rendering in the UI.
Resolution
Restart the job-launcher pod in the affected environment:
Identify the job-launcher pod:
kubectl get pods -n <namespace> | grep job-launcherRestart the pod:
kubectl delete pod <job-launcher-pod-name> -n <namespace>
Kubernetes will automatically recreate the pod. Once restarted, return to the Fusion UI and check the Jobs tab. The missing jobs should now be visible.
If the issue recurs, verify if a scheduled cleanup or cron job may be restarting the pod intermittently, and investigate further.