I have configured a port sensor to check a Windows server on tcp/445 every minute. When I look at the statistics & graph for that sensor, I often see periods of 60% or 80% downtime. When I look at the statistics table for each interval, I sometimes see two or more contigious periods where the 'Available' field is empty and 'Downtime' is some percentage, say 67%.
I don't understand how a port sensor can be anything other than 0% or 100% down in any interval, especually over more than one contigious interval. If there was no response to the tcp connection, hence the empty 'Available' field, how can if be 67% down?!