Network device monitoring
See every device, port and link next to the apps they carry
Discover devices across your address ranges, poll them over SNMP v2c or v3 with 174 built-in device profiles, and receive their traps, syslog and flows on collectors inside your network.
- 174 SNMP device profiles
- SNMP v2c · v3
- Traps · syslog · flows
How it works
From an address range to a device you can trust
Collectors do the polling and listening next to your devices. Your workspace stores the results beside your traces and logs.
- 01DiscoverSweep address ranges or add single devices, fingerprinted by sysObjectID and sysDescr.
- 02MatchThe most specific of 174 built-in device profiles wins, or Brain writes one when none fits.
- 03PollCollectors poll interfaces, health and hardware over SNMP v2c or v3, matched to a device profile.
- 04MapLLDP neighbours connect devices port to port on the topology map, with the evidence for each link.
- 05ListenThe same collectors receive traps and informs, syslog, and NetFlow, IPFIX or sFlow.
- 06AlertTurn on ready-made monitors for interface errors, link state, power and temperature, or write your own.
Device profiles
Monitored on the day it is found, even without a profile
Discovery picks the most specific built-in profile: an exact sysObjectID, then the longest prefix, then sysDescr. When nothing matches, Brain writes a profile from one sample walk of that kind of device, mapped to OpenTelemetry semantic conventions.
- Every device gets standard interface and system metrics, matched or not
- Brain grounds vendor OIDs in a shipped index of MIBs instead of guessing, within a spend cap per device kind
- A generated profile then covers every device of that kind
Topology
A map drawn from what your devices report
Each switch reports its LLDP neighbours port to port, and Tracix draws the map from those reports. Every link keeps the evidence it came from, and turns stale when nothing reports it any more.
Confirmed from both ends
A link both neighbours report scores higher than one only one side sees.
Evidence on every link
Each link lists the reports behind it and when they were last seen.
Gaps stay visible
A neighbour Tracix does not poll appears as a stub, so discovery gaps show.
Roles you can pin
Pin a switch as core, distribution or access when the map places it wrong.
Traps and syslog
Traps that open an alert and close it again
Every collector in a policy can listen for traps and informs, so devices send to the nearest one. A linkDown and its linkUp form one condition that opens on the first and resolves on the second, and the alert history keeps the outage.
- SNMP v1, v2c and v3 traps, plus v2c and v3 informs, on one listener per collector
- Repeats within 10 seconds are dropped, and a source sending more than 50 a second is rate limited with a warning
- Syslog in RFC 3164 or 5424 over UDP or TCP, with parsers for Cisco IOS, NX-OS and ASA, Juniper SRX, FortiGate and more
Active conditions · 2
Trap stream
Better together
A bad port, the change behind it and the app it slowed, in one place
Device metrics, traps and flows land in the same workspace as your traces and configuration history, so a problem can cite all three.
- Configuration managementBack up configurations, review changes as drafts and roll back, through the same collectors.Learn more
- Flow analysisTop talkers, conversation partners and raw records from NetFlow, IPFIX and sFlow.Learn more
- BrainAsk why a port is dropping frames and get the change, the trace and the device in one answer.Learn more
Device coverage
What each kind of device shows
Each device opens on a page laid out for its kind. The profiles named here are examples from the built-in library, not the full list.
| Device class | Example profiles | The device page shows |
|---|---|---|
| Switches and routers | Cisco Catalyst, Nexus, ASR and CSR; Arista; Juniper EX, QFX and MX; Aruba CX; Dell OS10; Huawei; MikroTik | Ports and optics, VLANs, MAC and ARP tables, routing table, link aggregation, OSPF |
| Firewalls | Palo Alto Networks; FortiGate; Cisco ASA and Firepower; Check Point; Juniper SRX; SonicWall; pfSense | Sessions and session rate, VPN tunnels, HA state, policies, virtual systems |
| Wireless | Cisco, Aruba, Ruckus and Meraki controllers; Aruba, Cisco, Fortinet and UniFi access points | Clients, access points up, channel utilization, radios, SSIDs, licenses |
| Load balancers | F5; Citrix NetScaler; A10 Thunder | Virtual servers, pools and members, connections, requests, expiring certificates |
| Servers and appliances | Dell iDRAC; HPE iLO; Cisco UCS; VMware ESXi; Cisco ISE | Load, filesystems, disk I/O, disks, management controller health |
| Storage | NetApp; Isilon; EMC Data Domain; HPE Nimble; Nutanix; Synology; TrueNAS | Capacity, volumes and pools, failed disks, IOPS, latency, replication lag |
| Power | APC; Eaton; CyberPower; Raritan; Server Technology; Tripp Lite; Vertiv | UPS battery charge, runtime and load; PDU power, current, phases and outlets |
| Environment sensors | APC NetBotz; Room Alert; Liebert | Temperature, humidity, dew point, contacts and alarms |
| Printers | Brother; Kyocera; Xerox; Zebra | Status, supplies, trays, covers and pages |
| Everything else | BlueCat, Infoblox, console servers, packet brokers, voice gateways, any SNMP agent | Interfaces, throughput, uptime and hardware |
The built-in library holds 174 device profiles, plus building blocks such as IF-MIB and LLDP-MIB that they extend. Every device gets standard interface and system metrics, whatever it matches. Profile definitions are derived from the Kentik SNMP profile library under the Apache License 2.0.
Which SNMP versions and ciphers are supported?
Does the map show the hosts behind each switch port?
Can Tracix collect streaming telemetry over gNMI?
What happens when no profile matches a device?
Start with one collector and one address range
Install a collector next to your devices, add an SNMP range, and confirm the devices it finds.