I’m facing occasional schema changes in source systems (like new/renamed columns or datatype updates) that sometimes break ETL pipelines. What’s the recommended way in Lumenore to handle schema changes gracefully and prevent downstream failures while keeping reports stable?