Thanks for the clarification. And yes, that last point is critical. Once the cluster comes back online, and available again, I need to understand how long it will likely take to synchronise back with the mirrored database before I can then manually fail over back.. I suppose it depends on the volume of transactions committed to the mirror while the cluster was offline, but I can do tests and simulations for that..

Where in the UI can you set the thresholds for the failover, or is it a T-SQL process only? Mind you, as I'm doing a manual failover, this doesn't reallly matter I guess..