Edit

Add web page knowledge sources in Azure SRE Agent

Add any publicly accessible web page by URL. Your agent automatically fetches and indexes the content so it can reference external documentation, status pages, and wiki articles during investigations.

Tip

  • Add any publicly accessible web page by URL. Your agent automatically fetches and indexes the content.
  • Reference external documentation, status pages, runbook sites, and wiki articles during investigations.
  • One-step setup: provide a URL, name, and optional description.

The problem: external docs stay external

Your team's knowledge isn't all in one place. Runbooks live on wiki sites. Vendor documentation is on external portals. Architecture diagrams and status pages are scattered across different URLs. When your agent investigates an issue, it can't reference these external resources - unless someone manually copies the content and uploads it as a file.

That manual process is tedious and creates stale copies. The original page gets updated, but the uploaded file doesn't.

How web page knowledge works

When you add a web page as a knowledge source, the agent fetches the page content and stores it for reference:

  1. You provide a URL, a name, and an optional description.
  2. The agent fetches the page content through an anonymous HTTP request.
  3. The page content is stored and indexed.
  4. Your agent can reference this content during conversations and investigations.

The fetch happens when you add the URL. Content is stored as a point-in-time snapshot of the page.

The agent fetches pages anonymously, without authentication credentials. Pages that require sign in, VPN access, or corporate SSO can't be indexed. To add content from protected pages, upload it as a file instead.

What gets indexed

Aspect Behavior
Content fetched Full page content from the single URL provided
Link following No. Only the specified URL is fetched, not linked pages
Authentication Anonymous. No credentials are sent with the request
Supported protocols HTTP and HTTPS
Fetch timeout 30 seconds
Refresh Manual. Delete and re-add the URL to get updated content

When to use web page knowledge

Web page knowledge sources work best for:

  • Public documentation: vendor docs, API references, cloud service guides
  • Status pages: service health dashboards, incident history pages
  • Wiki articles: publicly accessible knowledge base articles
  • Architecture overviews: publicly hosted architecture diagrams and design docs
  • Runbook sites: external runbook repositories accessible without authentication

For internal or protected content that requires authentication, use file uploads instead.

Limitations

Limitation Details
No authentication Can't access pages behind authentication, VPN, or SSO
Single page only Doesn't crawl or follow links to other pages
Point-in-time snapshot Content isn't automatically refreshed when the source page changes
30-second timeout Pages that take longer than 30 seconds to load fail
URL format Must be an absolute HTTP or HTTPS URL

Next step