Traceroute
Watch every router between our server and a website or IP appear live, with who runs it, where it is, and how long it takes.
Tracing the route
Sending the first probes. Hops appear below as each router answers.
- Target
- google.com
- IP address
- Resolving…
- Completed
- In progress…
Traced from our server. This is the path from our server, not from your device. Hover a latency to see each probe.
Raw output
Traceroute to google.com, 30 hops max
What is a traceroute?
When you open a website, your data doesn't travel in one jump. It passes through a chain of routers, run by different companies, until it reaches the destination. A traceroute makes that chain visible: every router on the way, the network behind it, and how long each step takes.
It's the tool to reach for when a site is slow or unreachable and you want to know whether the trouble is near you, in the middle, or at the other end. Ping tells you whether something answers. Traceroute tells you where along the way things change.
How this traceroute works
We send probes that run out of steps
Every packet carries a counter called TTL that each router lowers by one. The first probe starts with a TTL of 1, so the very first router drops it and reports back. The next probe starts at 2, then 3, and so on.
Each router gives itself away
When a router drops a probe it sends back a short "time exceeded" message from its own address. That address is one hop on the path, and the time it took to arrive is that hop's latency.
We find out who and where
For every address we look up its hostname, the network that runs it and a rough location, so you can see a path move from your provider to a transit network to the destination.
We read the pattern, not just the numbers
A slow hop isn't automatically a problem. We separate slowdowns that carry on down the path from routers that are merely slow to reply, and say which is which.
Each hop gets three probes, so you see three response times side by side. The trace ends when the destination answers, after 30 hops, or when several hops in a row go quiet, which is normally a firewall.
Is this my route? Not quite
The probes start on our server, so what you see is the route from here to the destination, not from your own connection. That's ideal for checking how a site or server is reached across the internet. To trace from your own device, open a terminal:
- Windows:
tracert example.com - macOS and Linux:
traceroute example.com
Only the outgoing path shows up in any traceroute. Replies can come back a different way, so a trace never shows the whole conversation.
How to read the results
Each row is one hop: its address, hostname, the network that runs it, a rough location, and the three response times. The colored bar on the left edge changes color when the path moves from one network to another, which is how you can see a hand-off from your provider to a transit network to the destination's own network.
- * * *
- The router didn't answer any of the three probes. Many routers are set to ignore or rate-limit traceroute, so a row of stars on its own says very little. It only matters if every hop after it is silent too.
- !H and !N
- A router told us it has no route to the host (!H) or to the network (!N). Unlike a star, that is a real answer, and it usually means the destination isn't reachable from there.
- !X
- A router or firewall refused the traffic on purpose (administratively prohibited). It's the clearest sign of a filter rather than a fault.
- Several addresses on one hop
- Traffic is being shared across more than one router at that step (load balancing). That's normal on big networks, and each probe may have taken a slightly different route.
Finding the real bottleneck
The common mistake is blaming the first slow hop. Routers treat traceroute replies as low priority, so one hop can look slow while traffic passes through it quickly. What matters is whether the slowdown continues. If latency jumps at hop 7 and every later hop stays at that level, that's the real point where the path gets long or congested. If hop 7 is slow and hop 8 is fast again, ignore it.
A big jump between two hops in different countries is usually just distance. Crossing an ocean costs tens of milliseconds, and no provider can change that. We flag persistent jumps and slow-only-here hops separately so you can tell them apart.
What a traceroute can't tell you
- The return path, which can be completely different from the way there.
- Anything inside a network that hides its routers, which some providers do.
- A precise location. Router locations come from IP geolocation and are approximate, which is why the map only plots hops with a known city.
- Who an IP address belongs to as a person. It shows networks, not people.
Traceroute, ping and DNS: which one do you need?
Start with an online ping test to see whether the host answers at all and how steadily. If it's slow or silent, traceroute shows where on the path things change. If the name won't even resolve, look at its records with DNS Lookup. To see what an address on the path belongs to, try the WHOIS Lookup or an IP lookup.
Frequently asked questions
What is a traceroute?
A traceroute lists the routers your data passes through on its way to a destination, and how long each step takes. It works by sending probes with a rising TTL and noting which router reports back at each step.
Is this the route from my computer?
No. The probes start on our server, so you're seeing the path from here to the destination. It's right for checking how a site or server is reached from the internet, but not for diagnosing your own connection. For that, run tracert (Windows) or traceroute (macOS and Linux) on your device. The commands are shown with each result.
Why does my traceroute stop before the destination?
Usually because a firewall near the destination drops traceroute probes. The website or server can be working perfectly well, and the trace just can't see the last steps. A TCP check on port 443 will tell you if the service itself is answering.
What does * * * mean in a traceroute?
That hop didn't reply to any of the three probes. Routers often ignore or rate-limit them while forwarding real traffic normally. If later hops answer, those silent routers are simply not replying and traffic is passing through fine.
Which hop is causing my high latency?
Look for the hop where latency jumps and then stays high on every hop after it. One slow hop followed by fast ones isn't a slowdown: it's a router answering probes slowly. The long-lived jump marks the real long link or congested point, and it often lines up with a change of network or country.
Is 100% packet loss on one hop a problem?
Not by itself. If the hops after it respond, the traffic is getting through and that router just doesn't reply. Loss that starts at a hop and continues all the way to the destination is the kind worth investigating.
What are the private addresses at the start of the trace?
Addresses like 192.168.x.x, 10.x.x.x and 172.16 to 172.31 belong to private networks. On a trace from your own device the first hops are usually your router and your provider's internal network. They only appear on our trace if our hosting provider uses private routing.
What is the difference between ping, traceroute and MTR?
Ping asks one question: does the destination answer, and how fast? Traceroute shows every router on the way. MTR runs a traceroute continuously and tracks loss and latency per hop over time, which is better for intermittent problems.
Can a traceroute tell me where a person is?
No. It only shows routers on the network path, and router locations are approximate anyway. To see what's known about an IP address, such as its provider and rough region, use an IP lookup instead.
Why does the path look different each time?
Large networks spread traffic across several routes, and routing changes as links are added or fail. Two traces a few minutes apart can legitimately differ, and the return path (which a traceroute can't show) may not match the outgoing one.