[AWS] More logging for restarting an instance in STOPPING state and fail fast for the restart

[Michaelvll]

A user encountered an issue where the AWS cluster takes a long time (>40 mins) to raise the no capacity issue when restarting a STOPPED cluster. This PR mitigates the problem by adding more logs when we are waiting an instance that is in STOPPING state and fail fast for the restart with capacity issue.

Tested (run the relevant ones):

• Code formatting: bash format.sh
• Any manual or new tests for this PR (please specify below)
  • sky launch --cloud aws --gpus V100:8 -i 0 echo hi; sky stop sky-8919-gcpuser
  • During the stopping of the instance: sky start sky-8919-gcpuser. We show more logging and error out if timeout
    
  • When the instance is stopped: sky start sky-8919-gcpuser. Show the capacity issue quickly.
• All smoke tests: pytest tests/test_smoke.py --aws
• Relevant individual smoke tests: pytest tests/test_smoke.py::test_fill_in_the_name
• Backward compatibility tests: bash tests/backward_comaptibility_tests.sh

[cblmemo]

LGTM except for several nits 🫡