How do I interpret MTR and Traceroute results?

MTR (My Traceroute, and originally called Matt's Traceroute) is a Linux/Unix utility that combines Ping and Traceroute functions into one easy-to-use program. WinMTR is a Windows clone of MTR. MTR is run on a computer to provide detailed hop test data (a series of brief measurements). MTR does not require installation or system admin rights.

MTR provides test results that are a combination of ping and Traceroute. Where Traceroute makes three ping tests to each hop, MTR continuously sends ping requests to each hop until you stop the test. For each hop in the path, MTR displays the:

There are options to copy or export the results as plain text or HTML.

Like a normal Traceroute, we are looking for something that replicates down the route list, through to the end. A single hop with high latency isn’t a problem—unless it’s the last one.

If a hop introduces latency, or packet loss, and that continues through the remaining hops, we have identified a problem.

In this first example, we see one hop with 23% packet loss, and one with a worst-case latency of 52ms. Neither of them is a problem since they don’t continue to the next hop in the list.

 

In another case, we see poor results starting at hop 11, and replicating through to the destination host. If the host was in the US and not Australia, this would be a concern.

 

One other thing to note is the variation in latency–jitter. There is a small amount of jitter in hop 12 (above), but it does not continue through. Even though we have higher latency in the final six hops, jitter is very good and not a concern.