Reconsider policy for failing to stop a mission

[andchiind]

Describe the improvement you would like to see
Currently we stop missions by moving to the stop state and then we call the robot interface to stop the robot. If we fail to stop the robot we keep trying for N number of times until we give up. Regardless of whether we give up or succeed, the same transition to idle occurs, only with different warning messages. I would like to change this policy so that we only try to stop the mission once (if we keep retrying then the original API call will time out, since we only give a response once we exit the stop state), and that if we fail to stop the mission we should go to either an Error state or Monitor state. If we reenter monitor then we should hopefully receive task statuses indicating that the mission either continued, or that all remaining tasks were cancelled.

How will this change existing functionality?
It will avoid API request timeouts when trying to stop missions, and it will prevent us from entering Idle in an inconsistent state where we are uncertain whether a mission is still ongoing on the robot.

How will this improvement affect the current Threat Model?
N/A