-
Feature Request
-
Resolution: Unresolved
-
L3 - Default
-
None
-
None
User Story (Required on creation):
Process instance modification in the UI is implemented as a token move. However, the semantics in the underlying API can differ from what users intuitively expect from a token move.
For example, if a token is in a called process instance and it is moved to the end event, then users expect the called process instance to complete and the calling instance to continue.
However, creating this modification in Cockpit means that first a new token spawns at the end event and completes immediately (since there is no further wait state) and then the existing activity is cancelled, leading to the cancellation of the entire called process instance.
Since it is not feasible to anticipate/simulate the effects of a planned modification (e.g. processes have business logic that influence sequence flow decisions), the next best thing is to educate our users as good as possible. In particular, there should be documentation that is understandable for Cockpit users (=> this is not the extensive API documentation at https://docs.camunda.org/manual/7.14/user-guide/process-engine/process-instance-modification/).
Functional Requirements (Required before implementation):
- The Cockpit documentation for process instance modification (https://docs.camunda.org/manual/7.14/webapps/cockpit/bpmn/process-instance-modification/) contains a section with example scenarios and shows what a certain modification would do
- The example scenarios cover common cases where people make mistakes based on our experience from support and forum
- Especially cases with call activities are covered
- The modification UI links to this section so that users have the possibility to educate themselves before making mistakes