Represent CPU graphs in 100 % not to max % during time interval of representation:
In instance view (server-address/provisioning/instances/) CPU graph and in server view (server-address/infrastructure/servers/) METRICS> CPU graph scale is from 0 to maximum value of % during the given time.
Setting representation graphs to 100 % view would be more convenient when comparing CPU usage on separate instances/VMs.
If one instance have no spikes it shows lets say 5% straight line and other instance which had a spike (lets say 95 % CPU usage) shows strait line at same 5% with one spike. Those graphs would be different and visually comparing them became difficult.