What causes "Failed to get MCP resource" in microsoft/autogen?

Inspect e.__cause__ — the true constructor error determines the fix.

microsoft/autogen · error · Error

Failed to get MCP resource

Error message

Failed to get MCP resource

What it means

Last-resort wrapper raised when SearchClient initialization throws anything that is not ResourceNotFoundError or HttpResponseError — for example DNS resolution failure, SSL certificate error, missing required constructor argument, or a TypeError from a wrong credential type reaching the SDK. The original exception is chained as __cause__.

Source

Thrown at python/packages/autogen-studio/frontend/src/components/views/mcp/api.ts:139

      throw new Error(data.message || "Failed to list MCP resources");
    }

    return data;
  }

  async getResource(serverParams: McpServerParams, uri: string) {
    const response = await fetch(`${this.getBaseUrl()}/mcp/resources/get`, {
      method: "POST",
      headers: this.getHeaders(),
      body: JSON.stringify({
        server_params: serverParams,
        uri: uri,
      }),
    });

    const data = await response.json();
    if (!response.ok) {
      throw new Error(data.message || "Failed to get MCP resource");
    }

    return data;
  }

  async listPrompts(serverParams: McpServerParams) {
    const response = await fetch(`${this.getBaseUrl()}/mcp/prompts/list`, {
      method: "POST",
      headers: this.getHeaders(),
      body: JSON.stringify({ server_params: serverParams }),
    });

    const data = await response.json();
    if (!response.ok) {
      throw new Error(data.message || "Failed to list MCP prompts");
    }

    return data;

View on GitHub (pinned to 027ecf0a37)

Solutions

Inspect e.__cause__ — the true constructor error determines the fix.
Normalize the endpoint to a full URL: "https://<service>.search.windows.net".
Validate all required config fields (endpoint, index_name, credential) are non-None and correctly typed before constructing the tool.
For TLS/proxy issues, configure the corporate CA bundle or exclude the search endpoint from interception.

Example fix

# before
config = AzureAISearchConfig(endpoint="mysvc.search.windows.net", ...)  # no scheme

# after
config = AzureAISearchConfig(endpoint="https://mysvc.search.windows.net", ...)

Defensive patterns

Strategy: try-catch

Validate before calling

from urllib.parse import urlparse

def endpoint_well_formed(endpoint: str) -> bool:
    p = urlparse(endpoint)
    return p.scheme == "https" and bool(p.netloc)

Try / catch

try:
    results = await tool.run(query)
except ValueError as e:
    cause = e.__cause__
    if "Unexpected error initializing" in str(e) and cause is not None:
        raise RuntimeError(f"Search client init failed: {cause!r}") from e
    raise

Prevention

Validate the endpoint URL (scheme + host) in your settings model.
Always inspect e.__cause__ for this wrapper — the real error is never in the wrapper text.

When it happens

Trigger: Calling _get_client/run() with a malformed endpoint (e.g. missing https:// scheme, plain hostname that fails DNS), or passing config values whose types the SDK rejects, producing an unexpected exception type inside the constructor.

Common situations: Endpoint configured without scheme ('mysvc.search.windows.net' instead of 'https://mysvc.search.windows.net'); corporate proxies/SSL interception breaking TLS; offline environments where DNS fails; None values slipping into config fields.

Related errors

AI-assisted analysis of microsoft/autogen@027ecf0a37 (2026-08-15). Data as JSON: /api/errors/c2096b1596cd9420. Report an issue: GitHub.