Create mirror queues and bucket

[hannes-ucsc]

Create the cloud infrastructure needed for mirroring, excluding any Lambda functions, most prominently the mirror queue (FIFO), the corresponding fail queue and the actual mirror bucket. The bucket should be non-public for now, matching the default access for all of our buckets. Ensure existing functionality isn't impacted by that infrastructure. Extend the manage_queues.py script to be able to handle the new queue. Resetting the indexer shouldn't affect the mirror queues.

[dsotirho-ucsc]

@hannes-ucsc: "The specifics of the S3 mirror bucket are still being discussed. It is clear that it will have to reside in a separate AWS account, but ownership and control of that account is yet to be determined. In lower deployments we will have to use a non-ODP bucket anyways. Assignee to implement this issue for lower deployments where the mirror bucket resides in the same account as the reset of the infrastructure and the bucket should be part of the main component. Whether mirroring is enabled or not should be configurable. By default, it should be disabled. Assignee to also update the prototype (#6857) to mirror to that new bucket."