# Migrations acknowledged to run with atomic=False and no CONCURRENTLY op (AtomicFalsePolicy).
#
# atomic=False drops per-migration transaction rollback safety, so a failure midway leaves
# partial changes committed and a bin/migrate retry re-runs the whole migration. That is only
# safe when the migration's operations are idempotent (a re-run is a no-op on already-applied
# rows). The usual reason to want it is a long-running data migration that commits in batches
# so it never holds locks or bloats WAL for one giant transaction.
#
# Before adding a line here, prefer keeping the migration atomic (the default) if the data
# volume is small, or moving a genuinely large backfill to an async migration.
#
# Adding a line means: I understand this migration runs without transaction rollback safety,
# and its operations are idempotent so a bin/migrate retry after a partial failure is safe.
#
# Format: one "<app_label>.<migration_name>" per line, e.g.
# posthog.1234_backfill_something_in_batches
posthog.1264_delete_revenue_analytics_user_product_list
product_analytics.0004_delete_revenue_analytics_insights
tasks.0079_backfill_task_repositories
