Update orchestration service to work with S3

[emmastephenson]

Action Requested

Update the orchestration service to work with S3 in addition to Postgres.

Acceptance Criteria

Given an S3 bucket and the appropriate connection information/authentication, the orchestration service can write eCR FHIR bundles to the S3 bucket. These bundles should be referenceable by the eICR ID, just as they are in Postgres.

Testing

Manual testing with S3 will be required. We likely don't need to keep an S3 bucket around for CI/CD testing, but we'll need to figure out how to mock one.

Background & Context

We're setting up the eCR Viewer to work on AWS, in order for the NBS team to use it, and writing to S3 will be needed.

Additional Information

We'll probably need to create an S3 bucket to ensure this works during local testing. Nick C. or Josh D. should be able to help provide AWS credentials to the team.

[emmastephenson]

Closing in favor of #1343.

[emmastephenson]

Re-opening: we're going to make orchestration work with S3 for now, and will come back later to make it happen via eCR viewer.

[emmastephenson]

Once this is integrated with orchestration properly, will reach out for a pair.