2
CLI backfill creates proper DAG runs for each logical date in the range, which keeps run history and any lineage or metadata tied to execution_date intact and queryable later, same as if those runs had happened on schedule.
Need to reprocess two weeks of a DAG after fixing a bug. Options are airflow dags backfill from the CLI, or manually clearing and re-triggering task instances in the UI. Does one preserve run history and lineage metadata better than the other?
CLI backfill creates proper DAG runs for each logical date in the range, which keeps run history and any lineage or metadata tied to execution_date intact and queryable later, same as if those runs had happened on schedule.
Sign in to reply.