I Benchmarked Just My Socks on 3 Nodes. The Numbers Were Fine — My Test Method Wasn't
I Benchmarked Just My Socks on 3 Nodes. The Numbers Were Fine — My Test Method Wasn't I spent an evening benchmarking my Just My Socks subscription across three nodes, and the most interesting thing I found wasn't in the results table. It was that my first three test runs were measuring the wrong thing entirely — and nothing in the output told me so. Here's the data first, then the trap, because…
On 15th November 2026, the author conducted an evening benchmarking session of their Just My Socks subscription across three nodes to test the performance. The results were initially surprising as the first three runs were measuring the wrong thing due to the benchmark script not going through the proxy. The script was written in Python and was using the Windows registry for proxy settings instead of the environment variables.
This resulted in identical outputs across three different nodes, which appeared correct at first glance. However, upon switching to a different node, the same test returned identical results, indicating that the proxy was not being used. The author identified two fixes for this issue: first, verifying the exit IP first before trusting the test results, and second, explicitly testing through the proxy using the command 'curl -x'.
This helped ensure that the proxy was indeed being used and that the test traffic was going through the correct node. It was also noted that latency is a snapshot and not a verdict, and that different nodes can affect the performance of certain services like Gemini. The author also mentioned that there is currently no column for packet-loss in the table, and if a clean way to measure per-node loss is found, it will be added.
The author maintains the justmysocks-guide project where the guide for this experiment can be found.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
