stephen-bracken opened a new issue, #62944: URL: https://github.com/apache/airflow/issues/62944
### Description The [probe urls](https://github.com/apache/airflow/blob/main/chart/templates/webserver/webserver-deployment.yaml#L203) for the api server and webserver templates default to `{webserver.base_url}/health` values in the Http probes, but this causes issues with multi cluster deployments. ### Use case/motivation We are rolling out airflow deployments that are running active-active on multiple kubernetes clusters using a shared database that is also active-active, but the liveness probes on the pods point to the webserver base url, which a common cname which we use for the webservers from both clusters. This means that if the primary webserver stops responding and fails the liveness / readiness probes then both webservers will register as unhealthy. To resolve this I would like to reroute the Http probes to `localhost/health` or the local pod url (e.g. `airflow-webserver-abcde.airflow-webserver.airflow.svc.cluster.local/health`) using the kubernetes dns. i.e. the webserver liveness probes point to `https://airflow-deployment.example.com/health`, which both webservers are using as the `base_url`. This would allow us to use a load balancer to route traffic to both webservers, whilst accurately marking webserver pods as unhealthy if they fail to respond to a local liveness probe. ### Related issues _No response_ ### Are you willing to submit a 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]
