IPV6 Support Webserver Gunicorn 0.0.0.0:8080

[decgowran]

Checks

• I have checked for existing issues.
• This report is about the User-Community Airflow Helm Chart.

Chart Version

8.9.0

Kubernetes Version

Client Version: v1.30.3
Kustomize Version: v5.0.4-0.20230601165947-6ce0bf390ce3
Server Version: v1.29.12-eks-2d5f260

Helm Version

version.BuildInfo{Version:"v3.15.3", GitCommit:"3bb50bbbdd9c946ba9989fbe4fb4104766302a64", GitTreeState:"clean", GoVersion:"go1.22.5"}

Description

By default gunicorn startup uses 0.0.0.0:8080 which in ipv6 only network on eks gives connection refused in health check and readiness probes.

need abilty to override startup with [::]8080 tp support ipv6 only networks as gunicorn does not support both startups

Relevant Logs

Defaulted container "airflow-web" out of: airflow-web, check-db (init), wait-for-db-migrations (init)
/home/airflow/.local/lib/python3.10/site-packages/flask_limiter/extension.py:337 UserWarning: Using the in-memory storage for tracking rate limits as no storage was explicitly specified. This is not recommended for production use. See: https://flask-limiter.readthedocs.io#configuring-a-storage-backend for documentation about configuring the storage backend.
/home/airflow/.local/lib/python3.10/site-packages/flask_limiter/extension.py:337 UserWarning: Using the in-memory storage for tracking rate limits as no storage was explicitly specified. This is not recommended for production use. See: https://flask-limiter.readthedocs.io#configuring-a-storage-backend for documentation about configuring the storage backend.
[2025-01-15 10:13:16 +0000] [39] [INFO] Starting gunicorn 22.0.0
[2025-01-15 10:13:17 +0000] [39] [INFO] Listening at: http://0.0.0.0:8080 (39)
[2025-01-15 10:13:17 +0000] [39] [INFO] Using worker: sync
[2025-01-15 10:13:17 +0000] [84] [INFO] Booting worker with pid: 84
[2025-01-15 10:13:17 +0000] [85] [INFO] Booting worker with pid: 85
[2025-01-15 10:13:17 +0000] [86] [INFO] Booting worker with pid: 86

Warning  Unhealthy  11m (x2646 over 17h)  kubelet  Readiness probe failed: Get "http://[2600:1f18:96c:2f04:15a7::b]:8080/health": dial tcp [2600:1f18:96c:2f04:15a7::b]:8080: connect: connection refused
  Normal   Pulled     6m8s (x178 over 17h)  kubelet  Container image "docker.getcollate.io/openmetadata/ingestion:1.5.10" already present on machine
  Warning  Unhealthy  41s (x1759 over 17h)  kubelet  Liveness probe failed: Get "http://[2600:1f18:96c:2f04:15a7::b]:8080/health": dial tcp [2600:1f18:96c:2f04:15a7::b]:8080: connect: connection refused

### Custom Helm Values

_No response_

[decgowran]

As a workaround Ive been able to edit deployment and override startup exec command with below after deployment of chart

airflow webserver --host [::]