UFM offers comprehensive diagnostics for your InfiniBand fabric, covering a range of categories:
Fabric configurations
Fabric topology
Hardware issues
Communication errors
Maintenance
Security
Switch module status
NVIDIA SHARP notifications
Events are notifications generated by UFM, indicating issues within the mentioned categories in the InfiniBand fabric. On the other hand, alerts are urgent notifications derived from events (many events can be configured as alarms based on customer preferences).
These detections are performed both before running applications and during standard operation. They help troubleshoot and notify network administrators of potential network issues before they escalate.
Events can originate from various sources:
SM traps
SHARP AM traps
UFM internal analysis, encompassing:
Internal detection of topology changes
Internal fabric analysis (based on IBDiagnet)
Internal monitoring of managed switches
Maintenance activities (device action tracking, licensing, cable integrity)
Threshold-crossing events determined by telemetry counter readings
| WebUI | REST API | |
| Events | UFM events can be viewed via the Events and Alarms WebUI view. Refer to Events & Alarms. | |
| For device-specific events, refer to the Events & Alarms. | N/A | |
| Configuration of events is managed within the Events Policy Tab in the Settings window | ||
| Alarms | UFM alarms can be viewed via the Events and Alarms WebUI view. Refer to Events & Alarms. | |
| Configuration of alarms is managed within the Events Policy Tab in the Settings window | N/A |
For showing all the UFM supported events, refer to Threshold-Crossing Events Reference.