I am running complex workflows in n8n where one workflow (the “server”) is triggered by another workflow (the “client”) using the MCP Client Tool and MCP Server Trigger nodes. My server workflow often requires significantly more time to complete than the current 1-minute timeout allows—sometimes up to 15 minutes or more.
Request:
Please add an option to configure the timeout for the MCP Client Tool and MCP Server Trigger nodes, so that users can set a value appropriate for their workflow’s needs.
Why I need this feature:
The current fixed timeout causes the MCP client to fail with a timeout error, even though the server workflow is still running and will eventually produce a result. This makes it impossible to use MCP for long-running processes, which are common in my use case.
How it would help:
Allowing the timeout to be configurable (either in the node settings or via environment variables) would enable workflows to wait for the actual completion of long-running tasks. This would make the MCP nodes much more flexible and suitable for a wider range of automation scenarios.
Setup details:
- Both the client and server are n8n workflows, running on separate n8n instances.
- The server workflow can take 15 minutes or more to complete.
- Both instances may be running behind a reverse proxy, but the main limitation is the hardcoded timeout in the MCP nodes.