If you’ve ever built analytical dashboards, chances are you’ve worked with the old “batch-load-to-RDS” pattern, where we load the data into RDS from a batch process and read from that operational data store, such as AWS RDS or other databases. We certainly did in our case — and for a while, it served us well.
Our applications pulled data from a lakehouse like Snowflake or Delta Lake, transformed it in AWS Glue, and loaded it into Amazon RDS. The RDS database powered Lambda APIs that fed dashboards and front-end apps.