Your Databricks job occasionally fails due to transient network issues. To ensure the job automatically retries before sending a failure alert, what configuration should you apply?