SameerMesiah97 opened a new pull request, #76958: URL: https://github.com/apache/airflow/pull/76958
**Description** This change adds `DeadlineReference.DAGRUN_START_DATE`, allowing DAG authors to calculate deadline alerts from when a DAG run enters the running state. The deadline is calculated as `DagRun.start_date + interval`. Start-date deadlines are created when the scheduler starts a queued run, or during creation if the run is created directly in the running state. **Rationale** DAG authors may need to monitor completion time from when a run starts, excluding its initial time in the queue. For example, a run queued at 09:00 and started at 09:10 with a 30-minute interval has a start-date deadline of 09:40. **Tests** Added and extended tests verifying that: - Start-date references evaluate the supplied DAG run’s `start_date` and add the configured interval. - An unset `start_date` returns `None`. - The public reference is registered in the appropriate categories. - The SDK reference serializes and reconstructs as the corresponding Core reference. - Runs created directly in the running state receive a start-date deadline. - Processing start-date alerts preserves existing queued-at deadlines. - Starting a queued run through the scheduler creates the expected deadline. - Subsequent scheduler passes do not create duplicate deadlines. **Documentation** Documented the new built-in reference and added an example showing how to configure a deadline relative to a DAG run’s start time. Added a feature newsfragment and updated relevant docstrings. **Backwards Compatibility** The new reference is opt-in. Existing deadline references retain their evaluation timing and callback behaviour. ###### Was generative AI tooling used to co-author this PR? - [X] Yes (please specify the tool below) Generated-by: [GPT 6.1] following [the guidelines](https://github.com/apache/airflow/blob/main/contributing-docs/05_pull_requests.rst#gen-ai-assisted-contributions) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
