Include resource sizing and timeouts in pipeline configuration.

As asked in https://github.com/developmentseed/aws-asdi/issues/18, when a pipeline is created, it can be integration tested locally via `docker run` with specific memory, cpu and storage constraints to determine initial resourcing and timeout baselines. These values can be included in the pipeline's `config.yaml` and used to dynamically set the granule processing Lambda's settings at pipeline deployment time.