potiuk commented on issue #33688: URL: https://github.com/apache/airflow/issues/33688#issuecomment-1691418941
From the discussion: HI @jaetma , Could you help answer the following questions that could help us check this more? 1. do you have some monitoring in place? Can you check what is the resource consumption for each of the Airflow components/pods? 2. Have you added any new DAGs that could be eating up a lot of resources and hence tasks are contending for resources? 3. What do the task logs say? Are they stuck at some specific step always? 4. How and where is Airflow deployed? 5. Which executor? Celery/Kubernetes/? 1. Currently, we only have this DAG running, with a maximum of 4 DAG runs, and those are taking all the machine resources. Tasks are taking over 2 minutes to start, and when running, they are also slow: 2. No, recently we have not added any DAGs with high resource consumption. 3. Logs are taking too much time to load in the Airflow UI. They are always stuck before starting, in the queued state. 4. Airflow is deployed on a server with docker-compose. 5. We use Celery as the executor.   Today we did a downgrade to 2.6.3 again, the issue seems to be resolved!  -- 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]
