Result
What does this mean?
- Loss
- Share of probe packets without a reply. What counts is the value at the target: up to 2 % is barely noticeable, from 10 % on video and calls start to stutter.
- Grey = harmless
- Loss at an intermediate hop that disappears at the following hops comes from the router’s rate limiting (ICMP rate limiting). Your data is not affected.
- Avg time
- Average round-trip time to this hop. A big jump between two countries is usually just the distance.
- Best / worst
- The fastest and slowest reply of all rounds. If they are far apart, the line is busy at times.
- Jitter
- How much the times vary from round to round. More than 30 ms at the target mostly affects calls and gaming.
- Network (AS)
- The network operator of a hop, looked up from public routing data. This shows in whose network a problem lies.
What the MTR analysis shows
MTR combines traceroute and ping: first the path to the target is found, then every hop receives one probe packet per round. After 10 or 20 rounds it is clear how reliably and how fast each hop answers – and where things get tight.
Our evaluation does the interpreting for you. It tells real loss that persists all the way to the target apart from harmless gaps at individual routers, and names the network where a problem begins. While the measurement runs, you see the interim results round by round.
When a measurement helps
Typical cases are a website that is only slow from Germany, a mail server that is unreachable at times, or choppy calls. You can copy the result with one click and send it to your provider or your IT team.
We measure from our test server in Germany and only the outbound path. Internal hops of our data centre are shown without their address; private and reserved addresses cannot be checked.
Frequently asked questions
One hop shows 40 % loss but the target 0 % – is that bad?
No. The router only answers probes at a limited rate but forwards your data completely. We mark such values in grey.
Why does MTR measure several rounds?
A single packet can get lost by chance. Only several rounds show whether loss happens regularly. 10 rounds are usually enough; 20 rounds are more precise for rare dropouts.
Does the measurement put load on the target?
Hardly. We send at most one small packet per hop per second and limit how often a target may be measured per hour and per day.
The target does not answer – what now?
Many servers filter probe packets. If the path into the target’s network is clean, everything is usually fine. Whether a website is reachable is shown by our ping test on port 443.
Domains and DNS in good hands
1blu business GmbH looks after domains, DNS and mail servers for companies, agencies and resellers – with dedicated contacts instead of a call centre.