Hello @Louis_Frolio !

this sounds very nice! I have already DAB implemented! A few questions to the above, but I think I more or less understand the rest. What is your suggestion in terms of keeping the version of migration or what do you mean by this? Is it about the prefix 001, 002 etc? Because I think we do not want to edit existing migrations so there will be always one version. In any change is required it should be done in a separate file:

001_create_base_schemas.sql
002_create_orders_tables.sql
003_seed_reference_data.sql
004_base_schemas_update.sql

Also currently we have yml files (001_create_base_schemas.yml, 002_create_orders_tables.yml etc.) separated per different tables creation and seed (exactly as you mentioned).

Shall we merge it to the one migration job and each of the migrations should be a separate task? 
Additionally every time when new migration file will be added shall we add new task to the migration yml? And in terms of "Your deployment job reads which versions are already applied" shall we check it in every file (at the top) whether it is already in the admin.schema_migrations? Could it be possible for you to give me a small draft how do you imagine this?

Thank you a milion in advance!