My Trace Route (mtr) Packet Loss Testing

When a connection feels flaky, the first question is always the same: where is the packet loss happening? mtr (My Trace Route) combines the functionality of traceroute and ping into a single tool, so you can see loss at every hop in one pass.

Running a Report-Mode Test

For a quick, scriptable capture, run mtr in report mode with a fixed number of cycles:

[root@server ~]# mtr -c 30 --report 62.113.4.14

Breaking that down:

• -c 30 sends 30 cycles of probes (one cycle = one ping to each hop).
• --report prints a static table and exits, which is ideal for logging or pasting into a ticket.
• 62.113.4.14 is the destination IP. You can swap in a hostname if DNS resolves.

Reading the Output

The report lists each hop with columns for Loss%, Snt (packets sent), Last, Avg, Best, Wrst, and StDev. A few things to watch for:

• Loss at a single hop only often means that router is rate-limiting ICMP responses. It does not necessarily indicate a real problem.
• Loss that appears at one hop and persists at every subsequent hop points to a genuine issue at or before that hop.
• Intermittent loss (1-5%) across multiple hops can indicate congestion or a misbehaving link.

Tips for Better Results

• Run the test during the time window when users report problems, not just at 2 a.m.
• Increase the cycle count (-c 100 or higher) for a more statistically meaningful sample.
• Compare results from two different source hosts to rule out a local NIC or switch issue.
• Pair the mtr output with ping -c 100 <destination> to confirm end-to-end loss independently of intermediate hops.

Quick Comparison: mtr vs. traceroute + ping

mtr is not a replacement for deep packet captures, but for 90% of “why is this slow” questions it gives you a clear, hop-by-hop answer in under a minute.

• Networking
• How-To

• Troubleshooting
• Bash