Server health
CPU, memory, network speed and last contact come from the agent. They describe the whole server, not the speed of one customer.
Where to find it: #/nodes-metrics — Infrastructure.
Workflow
Check measurement freshness. For high load, open the node and compare customer count, memory and traffic.
Stale measurements
Open the node and compare the last report with the current time. Check outbound HTTPS to the panel and sn-node logs. Interpret CPU and memory alongside customer counts and traffic; these are server measurements, not individual VPN-user measurements.
Verify the result
Zero activity does not prove a working connection. If metrics are stale, check the agent service and connectivity to the panel.