Update Job Deletion Logic To Support Force Kill

[d80tb7]

This changes the executor's job deletion logic so that:

• The first attempt to delete a job attempts to delete the job while respecting the graceful termination period set on the pod
• Any subsequent delete attempts that occur within the graceful termination period + killTimeout (killTimeout defaults to 5 minutes but is configurable) are no-ops.
• Any subsequent delete attempts that occur after graceful termination period + killTimeout attempts to delete the job with a graceful termination of 0 (i.e. force kills the job)

[d80tb7]

Here I want to assert that we using a timout of 0 to kill the pod, but it doesn't look like the fake client tells me this.

[JamesMurkin]

I had a look, I can't see how to get the options passed to the delete call either :(

Possibly google may have some ideas