fix(jupyterhub): use cluster DNS for internal Hub API calls

[idirze]

Description

Configure JupyterHub to use the Kubernetes fully qualified service DNS name for internal Hub communication via hub_connect_url.

This ensures spawned notebook pods receive a JUPYTERHUB_API_URL that uses the .svc.cluster.local DNS suffix, for example:

hub:
  config:
    JupyterHub:
      hub_connect_url: "http://jupyterhub-hub.<namespace>.svc.cluster.local:8081"

This allows internal Kubernetes service traffic to be excluded from the corporate proxy using a generic NO_PROXY / no_proxy suffix:

singleuser:
  extraEnv:
    NO_PROXY: "localhost,127.0.0.1,.svc,.svc.cluster.local,.cluster.local"
    no_proxy: "localhost,127.0.0.1,.svc,.svc.cluster.local,.cluster.local"

Without this, notebook pods may try to reach the internal Hub API through the corporate proxy and receive 403 Access Denied by proxy.

Related Issue

OKDP/trino-opal-example-policy#2

Type of Change

• Bug fix
• New feature
• Documentation update
• Refactor / chore
• Breaking change

How to Test

After deploying the chart change, restart the Hub and the affected notebook server.

Verify that a new single-user notebook pod receives the Kubernetes FQDN-based Hub API URL:

kubectl exec -n <namespace> -it <singleuser-pod> -- python -c 'import os; print(os.environ.get("JUPYTERHUB_API_URL"))'

Expected output:

http://jupyterhub-hub.<namespace>.svc.cluster.local:8081/hub/api

Verify that Requests does not select the corporate proxy for the Hub API URL:

kubectl exec -n <namespace> -it <singleuser-pod> -- python - <<'PY'
import os
import requests

url = os.environ["JUPYTERHUB_API_URL"]
print("JUPYTERHUB_API_URL =", url)
print("NO_PROXY =", os.environ.get("NO_PROXY") or os.environ.get("no_proxy"))
print("Selected proxies =", requests.utils.get_environ_proxies(url))
PY

Expected output:

Selected proxies = {}

Finally, verify that the notebook can call the Hub API without receiving a corporate proxy 403 response.