GB6638 opened a new issue, #55071: URL: https://github.com/apache/airflow/issues/55071
### Apache Airflow version 3.0.5 ### If "Other Airflow 2 version" selected, which one? _No response_ ### What happened? We are running 172 DAGs in Airflow. The UI (/dags page) takes around 15 seconds to load. After profiling network requests, we found that the slowest request is: `GET /airflow/ui/dags/recent_dag_runs?dag_runs_limit=1&limit=50&offset=0` Looking at the source code, the following line seems to be the main cause of slowness: `dags_with_recent_dag_runs = session.execute(dags_with_recent_dag_runs_select_filter)` The query dags_with_recent_dag_runs_select_filter takes a long time to run and seems to be inefficient. We did not experience this issue before on Airflow 2.9.1. It seems likely that changes in the UI in Airflow 3.0.4 caused the query to change, resulting in slower response times. Optimizing this query or caching recent DAG runs could significantly improve UI performance. We have already cleaned the metadata database using: `airflow db clean --clean-before-timestamp '2025-07-01' --skip-archive` However, even after this, the query still takes a long time when run manually. ### What you think should happen instead? Query Optimization: The dags_with_recent_dag_runs_select_filter query should be rewritten to fetch only the necessary fields and use indexes effectively. ### How to reproduce 1. Access the Airflow UI /dags page. 2. Observe the network request to /airflow/ui/dags/recent_dag_runs. Notice the slow response time (~10 seconds). Expected Behavior: The /dags UI should load faster even with a large number of DAGs. ### Operating System Ubuntu 22.04 ### Versions of Apache Airflow Providers Airflow 3.0.4 Python version: 3.10,14 ### Deployment Official Apache Airflow Helm Chart ### Deployment details Virtualenv installation ### Anything else? The bottleneck appears to be in the query dags_with_recent_dag_runs_select_filter. Optimizing this query or caching recent DAG runs could significantly improve UI performance. ### 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]
