Change queue behaviour to ensure things run if azure kills the container

[jordan-evens]

Right now the queue message is read and if the run isn't completely done with everything at the final check then a new message is added to the queue.

Seems like when azure times a job out and kills it then it ignores the retry count option (maybe the time is overall and not per attempt?)

Want to change it to add a message to queue at start of run and delete it if the run is complete, so that if the run is killed then there is a message in the queue that will retrigger it.