VladimirYushkevich opened a new issue, #34338: URL: https://github.com/apache/airflow/issues/34338
### Apache Airflow version Other Airflow 2 version (please specify below) ### What happened Following the upgrade of Airflow from version 2.6.2 to 2.7.0, the `webserver` began reporting errors related to the receipt of a SIGTERM signal from the worker. This issue was not observed in earlier versions. ### What you think should happen instead Here an example of log entry from `webserver`: ``` "name": "gunicorn.error", "filename": "glogging.py", "module": "glogging", "message": "Worker (pid:40) was sent SIGTERM!", "processName": "MainProcess", "timestamp": "2023-09-13 10:37:19,466", "severity": "ERROR" ``` Used K8s resources for CPU and Memory for all airflow pods are always ~50%. I also don't see any correlation with scheduled tasks(we don't have many tasks). We also don't have any CPU/Memory intensive tasks. ### How to reproduce Use the following in values.yaml when deploying the helm chart on GKE cluster: ``` executor: CeleryExecutor airflowVersion: "2.7.0" ``` ### Operating System Debian GNU/Linux 11 (bullseye) ### Versions of Apache Airflow Providers _No response_ ### Deployment Official Apache Airflow Helm Chart ### Deployment details _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]
