If you do not want to set event listener to every jobflow (in case you have many jobflows or you often create new jobflows), you can create graph which can check results of your jobflows. In your graph you can use server’s HTTP API, which has operation named executions_history. By this way you can get details about your execution history and you can react when any jobflow fails (send email). This graph can be started repeatedly on the server by Scheduling.
If you leave Job file field empty and set only a sandbox, the listener works for all graphs or jobflows within the sandbox. You can see this functionality used e.g. here: http://doc.cloveretl.com/documentation/ … ut-failure