A customer created a process instance and then they aborted it using the Rest API:
/server/containers/{containerId}/processes/instances/{processInstanceId}
 
They found that for the same application, one bpmn flow is working as expected (current node ) and the other bpmn flow is displayed as a stuck node with RED color and the process instance with an "aborted" stage. They're suspecting that the issue is reproducible on the flows which have an async event node.