This is an experiment to replace the batch processing with a Fargate task.
We should:
- create a new Fargate task in the AWS console
- name the fargate task and associated task something
-manual so it won't overlap
- get everything in place to run an analysis using the queue associated with this new task
- confirm that the analysis runs and works
- we may need a thin PR with changes to job definitions, but we don't need to PR if only changes to tfvars