This view performs a DBSCAN (clustering algorithm to group together similar points) to all the tasks of the selected job.
Both axis are configurable to show any metric, and tasks can be limited to a specific type or all at the same time.
Two parameters are required for DBSCAN: eps and minPoints. Eps is the minimum distance between two points to consider them as different clusters. MinPoints is the minimum number of points that are needed to form a cluster (will be marked as noise otherwise). If heuristic is enabled, it will calculate the best guess for both parameters considering the current points.