To split out values from the `issues` table and reduce indexes, we're decomposing the following values into a separate table: # Table: public.work_item_transitions | Column | Type | Collation | Nullable | Default | |--------|------|-----------|----------|---------| | work_item_id | bigint | | not null | | | moved_to_id | bigint | | | | | duplicated_to_id | bigint | | | | | promoted_to_epic_id | bigint | | | | As part of the work, we'll first 1. Create the table 2. Add a trigger to `issues` to sync the values to the table 3. Backfill the data 4. Start reading from the new table 5. Wait until the next required stop 6. Ignore the columns 7. Remove the original columns on the `issues` table