Online Traceroute
A real traceroute from a machine in the region you choose, hop by hop.
The route to trace
How this check works
Traceroute is a trick played on a safety mechanism. Every packet carries a hop counter so that lost packets cannot circle the internet forever; each router along the way decrements it, and the router that decrements it to zero sends back a note saying where the packet died. Send one packet with a limit of 1, another with 2, another with 3, and those notes sketch the entire path, router by router, with a round trip time to each. No browser and no edge worker can do this — it needs raw sockets — so this page hands the job to a Globalping probe in the region you pick, which runs the real command and returns every hop. Reading one is a skill worth thirty seconds: latency should step up gradually with distance, and what matters is a jump that persists to the end of the trace, not any single ugly number in the middle.
What the results mean
- Hop
- One router on the path. The first few belong to the probe's network, the middle to transit carriers, the last to the destination's host.
- Round trip per hop
- Time to that router and back. It normally climbs with distance; a sustained jump between two hops is where an ocean, or a congested link, was crossed.
- Silent hop (＊)
- A router that forwarded the packet but declined to send the note back. Traffic passes through it fine — answering traceroute is optional and many carriers switch it off.
- A slow hop that recovers
- One router showing 200 ms while later hops show 40 means that router deprioritises its own replies — routers answer diagnostics with their spare CPU. The path is fine; only that router's paperwork is slow.
Common problems and fixes
- The trace ends in stars before reaching my host
- If the site loads in a browser, the destination's firewall is dropping traceroute probes — extremely common, entirely benign. The trace still shows the path up to that point, which is usually the useful part.
- One hop shows terrible latency but the site feels fine
- Trust the end of the trace over the middle. Routers generate traceroute replies as a background task, and a busy router answers them late while forwarding real traffic at full speed. It only matters if every hop after it is slow too.
- The path looks absurd — my traffic goes through another continent
- Sometimes it genuinely does: carriers route where they have capacity and agreements, not where the map is shortest. If a route detours badly and the latency to that region matters to you, that is a finding for your host or CDN, who can often steer it.
- I need the route from my machine, not from a probe
- Run it locally: tracert example.com on Windows, traceroute example.com on macOS and Linux. This page shows how the world reaches your host; your terminal shows how you do. The comparison between the two is often the diagnosis.
Frequently asked questions
Why can't my browser run a traceroute itself?
Traceroute needs to set the hop counter on raw packets and listen for ICMP error replies, and browsers expose neither — deliberately, since raw sockets in web pages would be a gift to attackers. Any 'online traceroute' is therefore run from a server somewhere. This one is honest about where: a named probe in a region you chose, with the machine's network shown in the result.
What am I actually looking for in the output?
The place where latency jumps and stays jumped. A clean trace steps up gently with distance. A trace that reads 5, 8, 12, 110, 115, 118 crossed something expensive between hops three and four — an ocean, or a congested exchange. That location, not the final number, is the diagnostic.
Are the stars a problem?
Almost never. A star is a router that forwards traffic but does not answer diagnostics. Carriers disable those replies for load and policy reasons, so most long traces contain a few. The pattern that does mean something is stars from a certain point onwards forever — paired with a site that does not load, that is where the path actually breaks.
Why does the route change between runs?
Because the internet balances load across parallel links and reroutes around congestion continuously. Two traces a minute apart can lawfully take different paths to the same destination. That is resilience working. It is also why a single odd trace is a curiosity, and only a repeatable one is evidence.