Skip to content
koducathome

Traceroute Online

Trace the network path your data takes to reach a domain or IP address. Identify routing issues, latency, and packet loss using real network measurements.

How to use this tool

Steps for using the Traceroute Online.

  1. Type a domain name (such as github.com) or a public IP address (such as 8.8.8.8) into the box.
  2. The route is measured from the real Globalping probe nearest you — a web page can't send traceroute packets from your computer. The summary shows the probe's city, network and how far it is from you.
  3. Leave Protocol on TCP 443 to trace the way HTTPS traffic travels — it reaches most websites, including ones whose firewalls drop ping. Choose ICMP for a classic traceroute.
  4. Press Trace Route or Enter. The probe sends packets with increasing TTL to every hop and the table fills in as results arrive, usually within 5–20 seconds.
  5. Read the Route diagram and Hop Details table: each row is a router on the path with its packet loss, packets sent and last, average, best and worst round-trip times.
  6. Open Raw Output to see the probe's own MTR text, or copy the results and download them as TXT or JSON from Quick Actions.

Example

Tracing cloudflare.com from Mumbai

Target and location
cloudflare.com
Probe location: India
Raw MTR output from the probe
Host                                         Loss% Drop Rcv Avg  StDev  Javg
1. AS31898 _gateway (140.91.204.55)           0.0%    0   3 0.4    0.0   0.2
2. AS13335 162.158.226.47 (162.158.226.47)    0.0%    0   3 1.2    0.1   0.7
3. AS???   (waiting for reply)
4. AS13335 162.158.226.89 (162.158.226.89)    0.0%    0   3 1.8    0.4   0.6
5. AS13335 cloudflare.com (104.16.133.229)    0.0%    0   3 1.2    0.0   0.7

Common use cases

  • Find where latency or packet loss starts on the way to a website or API
  • Check whether a slow site is slow from one region or everywhere
  • See which networks (ASNs) and providers carry traffic to your server
  • Confirm a firewall or CDN change took effect on the network path
  • Collect a traceroute to attach to a support ticket with your host or ISP

What is traceroute and how does it work?

Traceroute shows the path packets take across the internet to reach a host: every router (a hop) in between, and how long each one takes to answer. It's the standard first step when a site is slow or unreachable from one place but fine from another.

It works with the IP TTL (time-to-live) field. The first packets are sent with TTL 1, so the first router discards them and replies with an ICMP time-exceeded message, revealing its address. TTL 2 reveals the second router, and so on, until the destination itself replies. This tool runs MTR, which repeats that several times per hop so you also get packet loss and best, average, worst and standard-deviation times.

Interpreting latency: times are round trips from the probe, so they normally grow as the path gets longer. A big jump that stays high on every later hop marks a slow or long-distance link — crossing an ocean adds tens of milliseconds by physics alone. A spike on one hop that later hops don't share is just that router answering probes slowly; it doesn't slow your traffic.

**Non-responding hops (* * *):** many routers are configured not to answer traceroute probes, or rate-limit them. A silent hop in the middle is normal if later hops still reply. If every hop after a certain point is silent and the destination never answers, the path is broken there or a firewall is dropping the probes.

Ping vs traceroute: ping measures the round trip to the destination only — it tells you whether the host answers and how fast. Traceroute shows where along the path delay or loss begins. Use ping to confirm a problem, then traceroute to locate it.

Troubleshooting with traceroute: compare traces from several locations; if the problem shows up only from one region, it's on a regional route rather than your server. Loss that starts at one hop and continues to the destination is real loss. Note the hop and its network (ASN) when you contact a hosting or network provider — and include the raw output.

Limitations

  • The trace runs from one Globalping probe, not from your own device or network, so your local path may differ.
  • Routers that ignore or rate-limit probes show as * * * or with loss even when traffic passes normally, and some destinations answer only one protocol — if one doesn't reach the target, retry with the other.
  • Only public domains and IP addresses can be traced — private, loopback, link-local, reserved and multicast addresses are refused.
  • Starting traces is rate-limited to keep the shared free quota available for everyone.

Frequently asked questions

Where is the traceroute run from?

From the real probe on the Globalping network (an open-source measurement platform by jsDelivr) closest to you. Browsers aren't allowed to send the low-level packets traceroute needs, so it can't run on your own computer; instead your browser shares your location (if you allow it — otherwise we estimate it from your IP address) and the trace starts from the nearest online probe — often in the same city or on the same network. The summary shows the probe's location and its distance from you. It is not run from your browser or your computer.

What does * * * mean in a traceroute?

That hop didn't reply to the probe's packets. Many routers are set to ignore or rate-limit traceroute traffic, so a silent hop is usually harmless as long as later hops — and the destination — still respond.

Why does a middle hop show packet loss but the destination shows none?

The router is deprioritising replies to probes aimed at itself, while still forwarding real traffic normally. Only loss that continues on every hop after it, through to the destination, indicates real packet loss.

Why is latency to a middle hop higher than to the destination?

Routers answer probes in software at low priority, so their own response can be slower than the time it takes them to forward packets. Judge the path by the trend and the destination's time, not one hop.

What's the difference between traceroute, tracert and MTR?

traceroute (Linux/macOS) and tracert (Windows) send a few probes per hop once. MTR combines traceroute and ping, sending several probes to every hop so you get loss and timing statistics. This tool runs MTR.

Why didn't the trace reach the destination?

Many servers — especially on AWS, Azure and behind CDNs — drop ICMP (ping) packets, so a classic traceroute stops short even though the site works. Tracing with TCP to port 443 looks like normal HTTPS traffic and usually gets all the way through. Use the Retry button to switch protocol.

Can I trace a private IP address or localhost?

No. The probe is on the public internet, so private, loopback, link-local and other reserved addresses aren't reachable from it and are refused. Domain names that resolve to such addresses are refused too.

What is sent to your server?

Only the target you enter and the protocol you pick, when you press Trace Route. If you allow location access, your browser's location (rounded to about 1 km) is sent so the trace starts from the probe nearest you; otherwise your IP address is looked up in a local database on our server for an approximate city. Neither is stored or sent anywhere else. Our server checks the target is public and asks Globalping to run the trace. Targets aren't stored or sent to analytics.