Reinforcement Learning Auto Scaling Demo

**Link to the notebook**
Add the link to the notebook.
https://github.com/aws/amazon-sagemaker-examples/blob/main/reinforcement_learning/rl_predictive_autoscaling_coach_customEnv/rl_predictive_autoscaling_coach_customEnv.ipynb

**What aspects of the notebook can be improved?**
It only plot rewards, but not actual scaling results, e.g. load vs capacity plot.
And actually the ppo preset performs very bad, is it possible to give a better tuned preset as starting point?

blue is load, red is capacity

**What are your suggestions?**

Start with reinforcement_learning/rl_predictive_autoscaling_coach_customEnv/rl_predictive_autoscaling_coach_customEnv.ipynb and run the existing reinforcement-learning example to inspect its reward plotting and PPO preset. Determine how the notebook represents load and capacity, then define the changes needed for a load-versus-capacity result and a better-tuned PPO starting preset. Done means both requested improvements are demonstrated in the notebook.

Written by the indexing model from the issue text.