1. Register one agent per node
A token identifies one reporting host, and a DaemonSet runs one pod per node, so register one agent per node in Monitor settings and name each after its node. Free plans cover three servers; a node is a server.
2. Put the tokens in one Secret, keyed by node name
kubectl -n realuptime create secret generic realuptime-agent-tokens --from-literal=node-a=rua_... --from-literal=node-b=rua_... The manifest mounts the Secret as files and sets REALUPTIME_TOKEN_FILE=/etc/realuptime/tokens/$(NODE_NAME), so each pod reads exactly its own node's token. A node with no key fails loudly rather than guessing.
3. Set the cluster label and apply the manifest
Edit REALUPTIME_CLUSTER in apps/agent/deploy/kubernetes/daemonset.yaml to your cluster's name and kubectl apply -f it. The pod mounts the node's root filesystem read-only at /host and reads the node's /proc, /sys and /run, so the sample describes the node, not the pod, and is recorded with vantage host.
4. Read the cluster view
The Server health section on the monitoring board groups hosts by cluster label with a roll-up line (hosts, average CPU, memory). Each node's page shows its containers, from cgroup v2, alongside everything else.
Go deeper
The full reference lives in the docs: Monitor agent documentation. Error codes named above are each explained in the error-code reference.