Add Terraform configuration for CloudWatch dashboards

[rbreslow]

Even though we're processing many studies in parallel, it looks like jobs are finishing sequentially. I believe that there's a bottleneck somewhere. It may be related to EC2 or database I/O.

This issue is to create a CloudWatch dashboard to visualize the flow of ETL -> Orthanc instance -> RDS instance to help us identify that bottleneck. Then, export that dashboard and codify it via Terraform.