There are four different elements that makeup a run, Triggers, Tasks, Errors, and Inputs.
When you first look at a Run, you get the Timeline view which is all elements in the order they were processed and the inputs. You can select the different elements to refine the view.
Inputs are static once the run has started. They normally take the form of a json string or XML document. Here is the input for a submitted request:
The source for the tree determines the input.
A trigger record tells the task engine that it needs to process the node (and following connectors). The most common trigger is a Start node. Besides start nodes triggers are used to complete a deferred node, fire Wait nodes, and during loop processing.
A task record is created for each node. It will show any outputs like deferred tokens, or compiled outputs. Depending on the type of task you can inspect the results or outputs. Here is the result of an Incident Create task.
An error record is created for every error after the failed trigger.
Errors are listed with newest first, but the best way to see related errors is from the error on the Run which has a link.
Example of a simple Run with an Error.
In the Error there is basic information, and a link to the specific error.
Clicking on the link to the error takes you to the actual error.
In this case the field on the Incident form has character limit that the data pushed from the task exceeded. After you fix this error you can return to this dialog and the Additional Action drop-down lets you Retry the error. You must enter notes about the fix, and then click the resolve error button.
The most important part of this dialog is the details panel. it provides as much information as possible to help you troubleshoot the error.