Many customers fail to see that the Usage graphs show averages between the day or week, and thus each point is not a particular case, but an average of many. Having a value too close to the limit could mean that the limit was touched then, but not for the entirety of that point (5 min, 15 min, 1h long). Since the lines are not touching on the graph, our customers get confused.

I think that if you plot the max / peak points instead of an average of values within each point in time it would be clearer when the resources get maxed out without looking particularly at the fault graph.

For instance if you have 20 EP in one minute, 10 on the next, 5 on the following three, an average graph with a 5 minute mark would show 9, whereas showing 20 (the max value) would be more useful.