How do you ensure the performance and reliability of the data pipelines you design in Azure Data Factory?