Nuances to config this properly

[rClark-qubeyond]

Thanks for the great helper function, though I've found two instances where this still may not work as expected, maybe something for the docs or a minor code tweak.

1. we happened to have tasks configured to use LogConfig awslogs for the container. By default, this driver uses 'mode":"blocking" which resulted in containers failing to relaunch and exiting with code 128 due to failure to reach the Cloudwatch services. Changing this to "mode":"non-blocking" in the task def seemed to resolve.

2. in another case, a multi-container task seems to exit on shutdown/reboot of the external instance to status "exited". This status apparently is not picked up with a restart policy of "on-failure". Changing the Python script to assign restart policy "always" in an offline scenario properly picked things up on reboot and reverted once online again.