1- What's going on with these packet-loss plots?
Maybe the readings are so close to zero in the first plot, and in the second tending to zero causing some glitch.
Possibly an auto-scale problem, maybe setting a minimum step could fix it?
2- Auto-scaling should not be enabled for the MCS plot.
There will never be a situation where it will be over 15 (picture 2)
Small variations can be exagerated and actually decrease readability of the graph (picture 4)
MCS 14,5? is that reading something like ["in the last 5 minutes i modulated 50% 14 and 50% 15"]??