I'm surprised that there aren't more comments on this, because it seems like it could be enormously useful. My first thoughts are that it could simplify synchronizing data between environments, or coordinating backups where a system has data in both S3 and RDS or RedShift.
Is it just me or does this look like AWS Data Pipeline with an actually feature-complete and usable user interface?
I just spoke one of the folks here who's PM for Glue. He confirmed that Glue will be superseding Data Pipeline, and it's basically the same team working on it.
Data Pipeline was a great version 1 of this idea, but the lack of functionality in the UI really killed it for me. It seemed inevitable that I would end up just writing the JSON configuration by hand.
I wonder how Segment, Fivetran, Stitch and all the other ETL vendors feel about this. I have a feeling that the ecosystem will flock towards Glue and leave everybody else in the dust...