is there a way to configure metrics in the MCP for new experiments? one of the last things that we do manually

let me ask the ai team