Queue overflow may cause this to happen.
Log Message:
… QueueFullException: monitoring.[CONNECTIONAME] queue length exceeded (200). Job rejected is [CONNECTIONAME].AcknowledgeAlarmsJob[2763606,MEDIUM,CANCELLED]
During queue overflow, collection data is lost. If the fail over occurs during a queue overflow, Spotlight will not collect this information. Once the service is restarted, the collections are restored and Spotlight then identifies the node changed host raising the alarm.
WORKAROUND:
Restart the Diagnostic Server services.
STATUS:
Waiting for fix in the future release of Spotlight on SQL Server Enterprise
