As the Operations Lead handling an incident with your service, you observe that a single node is returning 5xx errors for all requests while the service typically operates at 70% capacity. Customer support cases have also increased. To minimize user impact and adhere to Google-recommended practices, how should you proceed to remove the faulty node from the load balancer pool for isolation and investigation?