When I add a column to a stream, do I need to run a full refresh of that tap or does meltano recognize that new column and do a full refresh of that particular stream to whichever target I'm loading it to? How do you guys manage adding a new column to a stream if your tap/pipeline is running in a container in production? Do you like delete the table on the database side then have to go and run a full refresh for that tap any time you make a single column change to one of your streams?

The target would normally recognize the new column and update the destination table, but beyond that nothing else is bubbled up to Meltano and you need to manually run a full-refresh

Does running a full refresh tell meltano to truncate or recreate the target table or does that need to be done manually?

Not necessarily, depending on the loader's built-in or configurable strategy, it will either upsert, append or truncate the destination tables.

@Edgar Ramírez (Arch.dev) If I add a new column to a stream, say "customer_address", and I have a primary key in my existing downstream table with a primary key of "customer_id" which is populated in all rows for the full range of data already, and I pass a full refresh flag for the tap, will it upsert data such that it adds that new column to existing rows where customer_id = customer_id?