Two problems, one suggested solution: First, too often something important to us is hidden, because the graph showing the critical metric is skewed by another event. In order to show the peak of that event, the graph is scaled such that other activity is just a flat line at the bottom. Second, graphing multiple measures in a single graph, when their scales vary widely, makes interpretation difficult--for a contrived example, imagine plotting CPU (scales to 100) and batches per second (scales to tens of thousands for our web site). A tool for tuning MongoDB provides a beautiful solution, by allowing you to toggle between decimal and logarithmic scale. See the two graphs depicting the same period on this documentation page, in the "Logarithmic Scale" section, and notice the richness of information you can get from the log scale: https://rueckstiess.github.io/mtools/mplotqueries.html. I have used this to point developers to performance tuning opportunities that were otherwise hidden. I would love to see it in Redgate Monitor!