Are you using uptime robot to check http-get requests for node monitoring? Or are you monitoring the node port (default 28967)? You should monitor the node port, not the web interface. But either way, if the node ends up having continuous restarts, uptime robot might not catch it for a while.