Issue Summary

As described in Knowledge sources for Rovo agents, we are now capable of using specific knowledge containers as information sources to drive response quality with less ambiguity and more current data.

This feature is available for the majority of Rovo connectors and can be used to limit knowledge in a controlled way.

However, when using connectors for custom websites, there is no option to establish that boundary as expected. Currently, all kinds of webpages are covered under the Atlassian Public Documentation option and do not differentiate between content ingested from specific connectors or Atlassian public documentation pages.

On top of that there is also no option to restrict contents from the entire custom webpage, and use just a section of the external datasource as knowledge base. 

Steps to Reproduce

1. Configure a custom website Rovo connector
2. Create a Rovo Agent
3. Search the available options to use specific knowledge bases

Expected Results

Find an option to use the custom website connector in the list, then restrict the knowledge source based on specific sections/sites/parts of the custom website, based on URL patterns, or other filtering methods.

Actual Results

Adjust the configuration to use the Atlassian Public Documentation to ensure the Rovo Agent only uses data ingested from webpages.

Workaround

Currently, there is no official workaround for this behavior. When one becomes available, it will be added here. Rovo Agent instructions can be tailored to filter responses, ensuring they match a specific URL pattern, which helps limit the Rovo Agent's scope of knowledge.