Error in output of databricks managed mcp server

old_school — Sat, 01 Nov 2025 14:42:34 GMT

Hi Databricks Community,

Need suggestions.

I have added two functions and I am able to use them. I am using Cursor IDE but in the logs I can see below error,

2025-11-01 20:10:23.203 [info] Handling CreateClient action

2025-11-01 20:10:23.203 [info] Creating streamableHttp transport

2025-11-01 20:10:23.203 [info] Connecting to streamableHttp server

2025-11-01 20:10:23.213 [info] No stored tokens found

2025-11-01 20:10:23.784 [info] No stored tokens found

2025-11-01 20:10:24.324 [info] Successfully connected to streamableHttp server

2025-11-01 20:10:24.324 [info] Storing streamableHttp client

2025-11-01 20:10:24.324 [info] CreateClient completed, server stored: true

2025-11-01 20:10:24.325 [info] No stored tokens found

2025-11-01 20:10:24.344 [info] Handling ListOfferings action, server stored: true

2025-11-01 20:10:24.344 [info] Connected to streamableHttp server, fetching offerings

2025-11-01 20:10:24.346 [info] No stored tokens found

2025-11-01 20:10:24.966 [info] Found 2 tools, 0 prompts, and 0 resources

2025-11-01 20:10:25.016 [error] Client error for command Streamable HTTP error: Failed to open SSE stream: Not Found

mcp.json

"uc-fun": {

"type": "streamable-http",

"url": "https://my_url.cloud.databricks.com/api/2.0/mcp/functions/dev_sales/silver",

"headers": {

"Authorization": "Bearer <my_token>"

"note": "Databricks UC function"

}

Thank you.

Re: Error in output of databricks managed mcp server

AbhaySingh — Mon, 03 Nov 2025 16:06:50 GMT

The most important thing to know is: Databricks supports external MCP servers only if they use the Streamable HTTP transport mechanism. Client errors related to Streamable HTTP typically occur if the MCP server or client is not compliant with this transport protocol, or if the HTTP connection establishment fails.

topic Error in output of databricks managed mcp server in Generative AI

Error in output of databricks managed mcp server

Re: Error in output of databricks managed mcp server