abhijeets25012-tech opened a new pull request, #61494: URL: https://github.com/apache/airflow/pull/61494
### Description This PR adds documentation about running the API server with `server_type = uvicorn` on Kubernetes. Unlike `gunicorn`, uvicorn runs as a single long-lived process per pod and does not support rolling worker restarts. In long-running Kubernetes environments this can lead to gradual memory growth or stale internal state. The new section explains recommended operational practices such as Kubernetes rolling restarts, Helm-based restarts, and scheduled pod recycling. The goal is to help users operating Airflow on Kubernetes understand when periodic pod restarts are necessary and when switching to `gunicorn` may be a better production choice. ### What is included * New documentation section: **Running Uvicorn on Kubernetes** * Operational best practices for long-running deployments * Example `kubectl rollout restart` command * Guidance on uvicorn vs gunicorn behavior ### Why this change Users deploying with uvicorn on Kubernetes may assume automatic worker recycling similar to gunicorn. This clarification improves reliability guidance and reduces confusion for production setups. ### Testing * Documentation build checked locally * RST formatting verified * No functional code changes ### Notes This is a documentation-only change. -- 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]
