The whole query is below. Right now, let’s just focus on the secret sauce. The secret sauce is how DTU percentage gets calculated. In a nutshell, the maximum of CPU, Data IO, Log Write Percent determine your DTU percentage. What does this mean to you? Your max consumer limits you. So, you can be using 1% of your IO but still be slowed down because CPU could be your max consumer resource.
That’s a rather interesting finding. I think the next step (which may be so context-dependent that it’s not possible to generalize) might be to figure out what various workloads do to the metrics and if there’s a way to predict with some reasonable accuracy the expected DTU load given an anticipated change in workload, rather than seeing the value spike and reacting to it later.