Druzhba is a Python framework for building data pipelines that incrementally moves data from production (transactional) databases (currently PostgresSQL and MySQL) into analytical data warehouses (currently Redshift). With Druzhba, pipelines are defined by a directory of YAML files that specify the link between the source and target databases, including by listing the columns that should not be copied (unlike other tools that require users to enumerate all the columns that should be copied). Although Druzhba supports some in-flight transformations, it was designed as an ELT tool (since Seatgeek, which released Druzhba, uses dbt for data modeling and Luigi for more complex dependency graphs).