How Selector AI Collects SNMP Data

In this demo, Selector collects SNMP data from multivendor network devices through scheduled polling and real-time traps, then brings both together on one dashboard. The walkthrough covers four views:

• Device Health shows every device as a hexagon, grouped by vendor and colored by KPI status.
• Device detail shows the learned baseline for packet loss, the inventory Selector reads over SNMP, and CPU and memory readings against an 80 percent threshold.
• Active Alerts lists the traps devices send the moment something changes.
• The correlation graph links tunnel status, device health, and HA status at a single point in time.

The demo opens one device to show its learned baseline: no packet loss over the past 24 hours apart from two short drops that recovered on their own. Selector reads the hostname, IP address, operating system, vendor, model, and serial number over SNMP, adds type, site, and location, and describes every vendor’s devices in the same data model. CPU holds between 20 and 30 percent and memory around 40 percent, well below threshold. The demo then follows a trap that arrived at 10:30 reporting a VPN tunnel down on a firewall. The tunnel and port KPIs turn red, while device hard-down status shows ping and SNMP checks green throughout, confirming the firewall itself is healthy. The correlation graph shows what changed at 10:30, and the tunnel’s own timeline shows it had been down well before the alert with nothing else changing. The operator confirms it as an inactive tunnel and resolves the alert in minutes.