A company's data ingestion workflow, which uses Amazon SNS for notifications and AWS Lambda for processing, occasionally fails due to network issues, causing unprocessed data. What actions should be taken to ensure all data is ingested?