Cloud providers want to make sure they bill someone for every microsecond of compute, particularly GPU compute which is in demand. They constantly work towards good monitoring because their profits are directly on the line.