MySQL --> Redshift Incremental Backup

Is it possible to set up Redshift (or Lake Formation, or something else) to be a replica of a MySQL RDS with a reasonable delay? (minutes or hours)

It seems like your only options are:

• do a full backup/restore (via S3) which is going to take too long
• do this pipeline thing which requires your data to be timestamped and doesn't handle deletes https://docs.aws.amazon.com/datapipeline/latest/DeveloperGuide/dp-object-copyactivity.html

Over in the naked MySQL / SQL Server / PSQL world, it's been pretty easy since the pre-cloud days to take incremental backups (which MySQL does of course support) and restore them elsewhere (which is what I can't find in Redshift) or just do a read-replica / log-shipping thing

I feel like I must be just looking at the wrong products / features. Or does Redshift truly not support this and I should just use a regular ol' MySQL replica for BI-type queries where we don't want the latency of a full backup/restore? (or is there a different data lake or data warehouse product I should be using?)

Will AWS DMS (Data Migration Service - https://aws.amazon.com/dms/) meet your needs? It supports MySQL as a source and Redshift as a target. It has the capability to maintain database equivalence through ongoing replication. No tool is perfect and DMS has some limitations. A good write-up on these can be found here - https://medium.com/@eshant.sah/aws-dms-877bf50e1ec0