Mined Petri Net(5.9.1)

This visual reconstructs your real workflow from observed behavior, giving auditors a picture of how work truly flows rather than how it was meant to.

What you can conclude

• Where unexpected shortcuts and dead-end states exist in your process.
• Where a routine workaround has quietly taken hold, signaled by an unexpected direct path between statuses.
• Whether the recorded transitions follow your documented workflow: a discovered map that matches your documentation shows they do.

How this chart works

The diagram is inferred from thousands of observed status transitions, then drawn as a process map using process-mining techniques. Each circle is a state of the map, sized by how often issues visited it. Each box between two states is a move between them, colored by the role that most often makes it. Roles come from your role mapping. An account Jira records as an app, such as Automation for Jira, takes the bot role unless your mapping gives it another; any other account you have not mapped, whether a person, a service-desk customer or an account Jira returned no type for, shows as unknown. The arrows are all drawn in one gray at one width, so on this map neither an arrow's color nor its width encodes time or volume; Flow Performance draws the same map with each move shaded by how long work waited before it. An on-page assistant lets you ask about specific paths, role-by-role contributions, or differences between documented and actual workflow.

Reading the counts under the map

The line under the map accounts for every issue with activity in the selected window. Issues on the map are the ones whose path survived the rare-path filter and were drawn. Issues that never changed state produced no path: no status change in the window moved them from one place on the map to another, so there is nothing to draw — this includes issues whose only status changes stayed within one state of the map, and every status you have not yet grouped into a category shares one state, so grouping those statuses can reveal moves the map does not show yet. Issues dropped by the rare-path filter followed an end-to-end route rarer than the threshold you chose. Those three groups add up to the issues considered. The path-shape figure counts distinct routes seen before filtering, with the number of routes kept beside it, so a shape count that stays the same while the threshold moves is expected.