Today we don't have a good tool to track which datasets or dashboards are using most resources. Such information will greatly help us to proactively locate resource bottleneck and poorly-designed dashboards, optimize them to avoid capacity related outages.
This enhancement request is related to recent SEV-1 outage: https://support.oracle.com/epmos/faces/SrDetail?srNumber=3-38098685501&srProdClass=Cloud