Depends on how you have deployed the cluster in the first place…
With the autonomous operator you have to use DNS names, but this is all hidden away from you (for good reason - you will see in a minute…). To scale the cluster simply type kubectl edit cbc ${cluster_name} then edit /spec/servers/${index}/size and the operator will do the rest. Do this, it’s easier!
If you have rolled this cluster by hand, you can use the pod IP address kubectl get pod ${pod_name} -o wide and use the IP returned in the table, but this is not recommended as IP addresses in Kubernetes are not stable and can change, thus breaking things!! If you are using DNS which is stable, you probably already have a headless service referencing all pods in your cluster, this will create DNS entries in the form ${pod_name}.${service}.${namespace}.svc which you can use to address the new servers.