An update in Buildkite now allows users to retain the history of retried jobs, creating a new job record while preserving previous attempts. This enhancement enables users to see detailed information about when and why a job was retried, including whether it was initiated automatically or manually. Users can leverage environment variables to track the number of retries and specify conditions for automatic retries, such as triggering them based on specific exit codes or preventing them for crucial steps like deployments. This feature provides fine-grained control over job retries, as exemplified by Buildkite's internal use to automatically retry jobs affected by AWS Elastic Stack host shutdowns. For detailed guidance on configuring retries within the pipeline.yml steps, users are encouraged to refer to the command step documentation.