Beyond Wireshark: the traffic monitoring toolbox
▶ Start the interactive lesson
Monday 9:15. "The internet is slow for everyone at the Manchester office." You could start a capture… but on which of 40 switches? And slow compared to what? Sam: "Packet captures are a microscope. First you need the map and the thermometer. Start wide, then zoom in."
Start wide: graphs and counters
Interface counters (show interface) tell you about errors (CRC, input errors), discards (drops when a queue is full), and current utilisation. Rising CRC errors usually mean a bad cable, port or optic.
Who is using it: flow data
NetFlow / IPFIX / sFlow: routers and switches export summaries of conversations (who talked to whom, on which port, how much), without the full packets.
Testing the path
ping: is it reachable, and what's the round-trip time? traceroute / tracert: which hops does traffic take?
Zoom in: packet captures
tcpdump on Linux devices and firewalls, Wireshark on desktops, or the built-in capture on many firewalls and switches.
Key points
- Start wide: utilisation graphs and interface counters (errors, discards).
- Flow data (NetFlow/IPFIX/sFlow) = who is talking and how much.
- Path tools: ping, traceroute, mtr, iperf3 to find where it goes wrong.
- Packet captures (tcpdump/Wireshark) to see exactly what happens, used with permission and care.
Words you'll learn
CRC / input errors · Discards / output drops · SNMP monitoring · NetFlow / IPFIX / sFlow · mtr / pathping · iperf3 · tcpdump