I am not able create a dedicated DNS endpoint configuration.

[nabros]

Hello, I'm trying to access my dedicated endpoint, title_today_EP (Endpoint ID: 4479520322*****1600, dedicated domain: 447*****22722201600.asia-south1-849*****5214.prediction.vertexai.goog), but I'm encountering a DNS resolution error – I'm getting a "Could not resolve host" message when I try to ping or access the endpoint's domain.

I've confirmed in the Google Cloud Console that the endpoint is in an "Active" state, and the deployed model is showing as "Ready."

I've already tried various DNS troubleshooting steps, including flushing my DNS cache and using Google Public DNS, from multiple machines on different networks.

I've also waited several hours to allow for DNS propagation, but the issue persists. Could you please investigate why I'm unable to resolve the DNS for my dedicated endpoint?

[marckevin]

Hi @nabros,

Welcome to Google Cloud Community!

It seems the same post has been posted through a public issue tracker marked as resolved, with a modification on code: "Predict" was present in my code. It should have been "GenerateContent".

In addition, here are several suggestions on troubleshooting for future reference that could help resolve the issue on DNS resolution error: “Could not resolve host":

• Network Configuration: Ensure there are no issues with your network configuration:
  • Carefully check your firewall rules, which might be blocking access to the host. You can try temporarily disabling them to see if this resolves the issue.
  • Although you mentioned using Google Public DNS, recheck to ensure it is properly configured. Additionally, if you're using a DNS private zone, verify that your dedicated endpoint is properly configured.
  • If you're using a private network configuration, such as VPC Peering or Private Service Access, ensure it is set up correctly.
  • Finally, check your network connectivity to make sure there are no underlying issues causing the error.

• For best practice, re-validate your code and dedicated endpoint configuration, including re-checking your code and domain mapping, as small errors like spacing can cause DNS failures. Also, ensure there are no region mismatches when making a request, as this can cause network issues.

• Service Account Permission: Ensure your Service account that is trying to access the Vertex AI API has a necessary permission to access where the endpoint is deployed. Typically you need roles/aiplatform.user and roles/storage.objectViewer if your model uses a GCS bucket.

Was this helpful? If so, please accept this answer as “Solution”. If you need additional assistance, reply here within 2 business days and I’ll be happy to help.