Accessing Confluence Public Spaces with Read-Only Permissions

[OndrejV]

I have been working on setting up a Google Search agent using a Confluence connector. I've successfully created a Confluence application, connected it to Google Cloud and created a data store that has downloaded all the necessary documents.

The problem arises when trying to query the agent. Probably due to the data source access control, I can only receive results from a space where I have both read and write permissions. I am not able to query documents from public spaces, with read-only permissions.

I have followed the provided documentation  (https://cloud.google.com/generative-ai-app-builder/docs/connect-third-party-data-source#confluence ) and ensured that the data store includes documents from all relevant spaces (it has identified over 1,000 documents, even though the space I can access only contains about 10 documents). The Confluence app has the correct permission scopes, and my email visibility is set to "Anyone."

I would greatly appreciate any insights or advice. Thanks in advance for your help.

[MJane]

Hi OndrejV , 

Welcome to Google Cloud Community!

The issue you encountered with accessing documents from public spaces with read-only permissions is likely caused by the data source access control feature within Vertex AI Agent Builder.

Here are possible workarounds you can try that might help you resolve the issue: 

Access Control Settings - Ensure that the access control settings for the data source are correctly configured to allow read access to public spaces. You might need to adjust the permissions to ensure that the agent can access documents from spaces with read-only permissions.

Service Account Permissions - Ensure the service account linked to the connector has the required access rights for the public spaces. It might be necessary to assign additional roles or permissions to this service account.

Verify Confluence App Permissions -  Double-check that your Confluence app has the necessary permission scopes.

Test with a Different Account -  Try accessing the agent with a different Google Cloud account that has write permissions to the public spaces.

If the issue persists, I suggest contacting Google Cloud Support as they can provide more insights to see if the behavior you've encountered is a known issue or specific to your project.

I hope the above information is helpful.