Packet loss is difficult to debug because it rarely fails in a consistent or obvious way. You might run a quick ping and see no issues, while your application still experiences timeouts or sporadic failures. From the outside, the network appears stable, but something along the path is clearly degrading the traffic.
The core problem is that common tools don’t fully answer where the issue occurs. Ping only tells you whether packets reach the destination. Traceroute shows the path they take. MTR provides more detailed data, but interpreting it. especially when the loss is sporadic, often requires manual analysis and experience.
In real-world scenarios, packet loss is usually conditional. It may appear only under load, affect a specific segment of the route, or occur in short bursts. By the time you begin investigating, the issue may no longer be visible. This makes both debugging and documenting the problem unnecessarily difficult.
What’s often missing is a clear way to identify where packet loss begins.
Ploss is built to focus on that problem. It runs continuous probes, analyzes packet loss across each hop, and highlights where loss starts and how it propagates through the path. Instead of just presenting raw data, it aims to make the structure of the problem visible.
A typical run looks like this:
<code>ploss 8.8.8.8 --mtr</code>
The output makes it easier to see which hop introduces loss and whether that loss continues downstream. When earlier hops remain stable and a later hop shows consistent loss, it becomes much easier to isolate the responsible segment, whether that’s within a local network, an ISP, or further upstream.
This is particularly useful for sporadic issues. By observing behavior over time, you can detect patterns that a single snapshot would miss. When escalation is necessary, exporting results as JSON or CSV provides a structured way to document and share findings.
Currently, Ploss supports per-hop loss detection using MTR, basic bottleneck identification, structured exports, and a watch mode for continuous monitoring. It runs on macOS and Linux and is designed to stay simple to use from the command line.
Planned improvements include multi-target probing, cloud endpoint testing, and an agent mode for continuous monitoring.
The project is available here:
<a href="https://github.com/Devjosef/Ploss" target="_blank" rel="noopener nofollow">https://github.com/Devjosef/Ploss</a>

Josef Röyem

Packet loss is notoriously hard to debug because it's often sporadic, conditional, and may disappear by the time you investigate. Common tools like ping and traceroute only provide partial answers. Ploss is an open-source CLI tool that runs continuous probes, analyzes loss per hop using MTR, and highlights where loss originates and how it propagates. It supports per-hop loss detection, basic bottleneck identification, structured JSON/CSV exports, and a watch mode for continuous monitoring on macOS and Linux. Planned features include multi-target probing, cloud endpoint testing, and an agent mode.

Debugging packet loss without guessing