On task detail, if the task is FAILED, allow leftover data deletion. Also ensure that the working folder of a task is not deleted if the task status reaches FAILED.
Instead of leftover data deletion of failed tasks implement deletion of all leftover data when the task/context/benchmark is removed from cluster, no matter if it is in state FINISHED or FAILED.
On task detail, if the task is FAILED, allow leftover data deletion. Also ensure that the working folder of a task is not deleted if the task status reaches FAILED.