We found the issue with the one database server and have fixed it. The overall performance of the system (Database, API and Delivery) has gone back to normal.