Instead of one large batch of ids, support single ids (keep batch functionality if possible) to allow job streaming. Completed jobs should insert their data into the database immediately after completion, resulting in nearly lossless runtime interruption of service.
Instead of one large batch of ids, support single ids (keep batch functionality if possible) to allow job streaming. Completed jobs should insert their data into the database immediately after completion, resulting in nearly lossless runtime interruption of service.