The graph we have for latency to a host is logarithmic. We have a risk that people will not appreciate the differences and miss potential faults. In the attached example 0-1 second response time, is the same as 1-10 seconds, and the jumps from 1-2...