MTR Online (My Traceroute)
Combined traceroute and ping — continuous hop stats with packet loss per hop.
About MTR
Combines traceroute and ping into one continuous test, running repeated probes to every hop along the path and aggregating loss and latency statistics (average, best, worst, jitter) per hop rather than a single snapshot. A single traceroute only shows one moment in time — MTR's continuous sampling is what actually catches an intermittent packet loss on one specific hop that a one-off ping or traceroute would never surface. Jitter is often the more telling number for real-time applications like VoIP or gaming, where consistency matters more than raw speed.
Packet loss shown at an intermediate hop doesn't necessarily mean real loss on that path — many routers deprioritize their own ICMP responses used for traceroute/MTR probes, so loss that clears by the final hop is usually an artifact of that hop's handling, not a real problem.
Column Guide
Avg Average RTT (ms)
Best Best RTT seen
Worst Worst RTT seen
StDev Jitter / stability
Related Tools
Reading the results
Each row is one hop between our server and the target. For every hop you get packet loss and latency (average, best, worst) plus jitter, the variation between probes. Two patterns separate real problems from noise:
- Loss that continues to the end is real. If loss starts at a hop and every hop after it, including the destination, shows similar loss, packets are being lost there.
- Loss at one hop only is not. If a middle hop shows loss but the destination shows 0%, that router is simply rate-limiting its replies to probes.
- The same applies to latency: a jump that stays high on every following hop is real, often a long-distance link. A jump on one hop only is a slow reply from that router.
Hops with no reply
Some routers never answer probes and show as blank or ???. That is harmless as long as later hops and the destination answer.
Direction matters
This test runs from our server towards the target. The return path can be different (asymmetric routing), so a problem may only show up in one direction. When you report a problem to an ISP or hosting provider, run MTR from both ends and include the reports:
- Linux / macOS:
mtr -rwbzc 100 example.com(report mode, wide output, IPs and names, AS numbers, 100 probes). - Windows:
pathping example.com, or WinMTR.
To see which network a problem hop belongs to, look up its IP with the ASN lookup.