Underlying code that raises this alarm is only run when the connection is first made to the SQL Server. It is not rechecked periodically (as most of the “normal” alarms are). So, this alarm won’t be cleared unless another failover occurs, or user turns monitoring off/on for that connection.
To have the error clear, an alarm action can be used. Here’s an example of conditions and actions:
For all alarms
where the alarm has existed for more than 2 hours
and the alarm is Clusters - Failed over
decrease the alarm severity to Normal
Enhancement request has been submitted for this feature to be included in a future release.
Enhancement SOSSE-4067 submitted to allow some alarms to block turning off “requires acknowledgement” and enhancement SOSSE-4068 to create an alarm action to “automatically acknowledge” alarms.