Differentiate timeout vs idle timeout

[TheRealHaoLiu]

currently idle timeout and timeout result in the same status call back this result in the AWX not being able to parse the differences between the two failure condition and make it confusing to debug

[relrod]

This will need a default like timed_out has on line 37. Otherwise if we don't hit this case, we'll crash when we check the variable below.

[Shrews]

You are missing checks in several places where currently self.timed_out is checked for cleanup, return codes, etc.

Also, we might need to consider if this is a breaking change because status value is now different, which is sort of an API change. Not sure this can be backported.

[AlanCoding]

Really quick 2 cents here - I also think the API change of a new status would be disruptive and prone to break things. Not something we want in the error handling! If you look a layer deeper in the call stack, other options present themselves.

ansible-runner/ansible_runner/runner.py

Lines 101 to 109 in d392c51

	def status_callback(self, status):
	self.status = status
	status_data = {'status': status, 'runner_ident': str(self.config.ident)}
	if status == 'starting':
	status_data.update({'command': self.config.command, 'env': self.config.env, 'cwd': self.config.cwd})
	for plugin in ansible_runner.plugins:
	ansible_runner.plugins[plugin].status_handler(self.config, status_data)
	if self.status_handler is not None:
	self.status_handler(status_data, runner_config=self.config)

Instead of using a new status value for the call to status_callback, it would go much more smoothly to add more data to status_data in the status_handler. I think you could even safely abuse job_explanation, because this failure path is a known scenario and it shouldn't be conflicting with anything else setting that string. This could relatively nice present the timeout reason to the user.