NetKit

Traceroute online

Understand traceroute in two minutes with the interactive TTL demo, then paste your own traceroute, tracert or mtr output to chart latency hop by hop and find where the delay really starts.

How traceroute finds each router

Every IP packet carries a Time To Live counter. Traceroute sends probes with TTL 1, then 2, then 3… and each router that kills a probe reports back. Step through it:

TTL = 1. Each router decrements the TTL by one. At Home router it hits zero, so that router drops the packet and sends back an ICMP "Time Exceeded" message, revealing its address 192.168.1.1. Traceroute times the reply: 1.2 ms.

Traceroute analyser

Run traceroute example.com (macOS/Linux), tracert example.com (Windows) or mtr -rwc 20 example.com, then paste the output. Parsing happens in your browser.

Paste output

Hops appear here as a latency chart once you paste output.

What a healthy traceroute looks like

  • Latency climbs in steps: under 5 ms inside your home, 5–20 ms to your ISP, then jumps that track distance.
  • Router names often contain location codes (lon, fra, nyc, iad) and link types (ae- for aggregated Ethernet).
  • The last hop's time is what matters. That is the real round trip, and it should match a ping to the same host.

Questions people ask

Why can't this site run a live traceroute?
Traceroute needs to send packets with a hand-set TTL and receive ICMP "Time Exceeded" replies, which requires raw sockets. Browsers and serverless cloud functions (where NetKit runs) don't allow them. A traceroute from a data centre also wouldn't show your path, which is usually what you need. Run it on your machine and paste the output here.
What do the * * * lines mean?
That hop didn't reply within the timeout. Many routers are configured not to send ICMP Time Exceeded, or rate-limit it. If later hops reply normally, the stars are harmless.
Latency jumps at one hop and then drops again. Is that a problem?
No. Routers answer traceroute probes on their slow control plane, at low priority. A spike that does not persist to the following hops is just that router being slow to reply, not slow forwarding. Only a jump that carries through to the destination indicates real added latency.
What is the difference between traceroute, tracert and mtr?
traceroute (Linux/macOS) sends UDP probes by default; tracert (Windows) uses ICMP; mtr combines traceroute and ping, repeating probes to show loss per hop over time. mtr is the best tool for intermittent problems. All their output can be pasted here.
How do I use TCP traceroute?
When UDP/ICMP is filtered, use sudo traceroute -T -p 443 host (Linux) or tcptraceroute host 443. On Windows, pathping gives per-hop loss statistics.