Hello guys! I’m setting the Datahub integration with Okta to configure SSO but I’m facing some errors during the deployment using Helm. Currently, I’m using the Datahub version 0.8.31. My current configuration into values.yaml:

extraEnvs:
  - name: AUTH_OIDC_ENABLED
   value: "true"
  - name: AUTH_OIDC_CLIENT_ID
   value: "<redacted>"
  - name: AUTH_OIDC_CLIENT_SECRET
   valueFrom:
    secretKeyRef:
     name: datahub-oidc-secret
     key: datahub-oidc-key
  - name: AUTH_OIDC_DISCOVERY_URI
   value: <identity_provider_host>/.well-known/openid-configuration/
  - name: AUTH_OIDC_BASE_URL
   value: <my_datahub_host>
  - name: AUTH_OIDC_SCOPE
   value: "openid profile email"
  - name: AUTH_OIDC_USER_NAME_CLAIM
   value: "email"
  - name: AUTH_OIDC_USER_NAME_CLAIM_REGEX
   value: "([^@]+)"

All pods are fine, but when I try to access the Datahub frontend using the host configured into Ingress, I receive an internal server error (500). When I check the logs from datahub-frontend pod, I found a connection reset error. 😕 Reading the documentation, I found this configuration here: https://datahubproject.io/docs/deploy/aws/#expose-endpoints-using-a-load-balancer But, my Kubernetes cluster already have an ingress controller configured. My question is: this step (expose endpoint using a load balancer) is mandatory or I’m able to reuse the configured ingress controller from my Kubernetes cluster? Thanks for help!

I replaced the default load balancer specified in helm with clusterIP. Then pointed an external ingress not specified at helm at the service. Worked for me.

Thanks! I will try these adjustments in my deployment. 🙂