In a scenario where you are tasked with managing a batch load process in Azure Data Factory, and you encounter a failure during the batch load, how would you handle the situation to ensure data integrity and provide a rollback mechanism?