There might be cases where jobs executed by jbpm executor get stuck in running state - best example of this is when executor is processing long running job and server goes down, the job is marked as Running in data store but the job did not finish so it won't be picked up by next executor run. To avoid such situations there should be possibility to dequeue given job to be picked up again.Two options should be available:

• automatic option to be triggered on server/executor start to find jobs that are in running state longer then defined time
• manual option to dequeue single job identified by id