Cancel a running detached job

Stops the job, terminating its WHOLE process tree (a training run is rarely a single process) and releasing any GPU it had reserved. Cancelling a job that has already finished is not an error — you simply get its final state back. Cancellation is not reversible: work since the last checkpoint is lost, so confirm with the user before cancelling something long-running. For a job that is still alive the reply is not immediate: the endpoint waits a few seconds for the process tree to actually go, so that it reports what became of the job instead of its state before the signal was sent. Allow a client timeout of at least 10 seconds. If the job outlives that wait the reply says so explicitly — the cancellation was still accepted, so do not repeat it.