How do you optimize a slow-performing SQL query when dealing with millions of rows of supply chain data at Jabil?