Do recognized generic guidelines exist for CPU utilization metrics from system performance testing of embedded real time applications? What is the safe range? What is the unsafe range? Should I be concerned if a supplier is delivering a system that includes a utility, servicing a key real-time data source, having a CPU utilization metric in the range 60% to 70%?
Software/Hardware used: Source data supplied in real time 24/7 into a distributed real-time system with public facing online web HCI that also interacts with external agencies.