Sure. We're breaking our metrics down into groups.
First we start with all of our documents in one group. In your screenshot above we're asking for maximum RxPeakKbps, so it's going to give us the maximum overall. When we scroll down and add bucket aggregations, we're splitting these metrics up into more groups.
So we'll say make groups of equal size over our time range, and this'll give us an x value (time range), and y (max RxPeakKbps for this time range). Splitting the series will take each group(time range) and split it further into groups by hostname, giving us two RxPeakKbps to plot.
It'll look something like this:
