je-enf opened a new issue, #73240: URL: https://github.com/apache/airflow/issues/73240
### Under which category would you file this issue? Airflow Core ### Apache Airflow version 3.2.1 ### What happened and how to reproduce it? Part of our pipelines behavior relies on the datetime of the last successful execution of a given task. To do that, we call the API endpoint **api/v2/dags/<dag_id>/dagRuns/~/taskInstances?limit=1&state=success&order_by=-logical_date** This mostly works as expected, however we randomly get the wrong result and I can't figure out the reason so far. For instance, we have a DAG scheduled once a day, and for a given task, the values returned by the API for the last 5 days have been: - **2026-09-12** : returns **2026-09-11** > OK - **2026-09-13** : returns **2026-09-12** > OK - **2026-09-14** : returns **2026-09-13** > OK - **2026-09-15** : returns **2026-09-14** > OK - **2026-09-16** : returns **2026-01-01** > NOK If I clear the task, then the second run usually get it right. ### What you think should happen instead? Behavior of the API should be consistent. ### Operating System debian ### Deployment Official Apache Airflow Helm Chart ### Apache Airflow Provider(s) _No response_ ### Versions of Apache Airflow Providers _No response_ ### Official Helm Chart version 1.21.0 ### Kubernetes Version _No response_ ### Helm Chart configuration _No response_ ### Docker Image customizations _No response_ ### Anything else? _No response_ ### Are you willing to submit PR? - [ ] Yes I am willing to submit a PR! ### Code of Conduct - [x] I agree to follow this project's [Code of Conduct](https://github.com/apache/airflow/blob/main/CODE_OF_CONDUCT.md) -- 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]
