What

The manager.go periodically checks the health of the MCP Server it is managing. It does this on a set interval. This means on failure it waiting the set interval before retrying. A better approach maybe to switch to an exponential backoff on error