Measure and visualize your internet speed systematically
Make sure you have Python3 and pip installed, then run this:
$ python3 -m pip install --user git+https://github.com/AdrianVollmer/speedypy
On Linux, put something like this in your crontab with crontab -e:
17 * * * * /home/<USERNAME>/.local/bin/speedypy --log-level ERROR measure
This runs the measure subcommand once an hour, at 17 minutes after the
full hour. One measurement per hour is a sensible frequency.
To visualize, run speedypy plot time_series.
Read the output of speedypy -h for more information.
We adhere to the XDG Base Directory
Specification,
so the configuration file is in $HOME/.config/speedypy by default.
In $HOME/.config/speedypy/speedypy.conf you can exclude certain servers
that consistently yield outliers. It could look like this:
[DEFAULT]
exclude_servers = 1338,16633,15728
You'll find the server IDs in $HOME/.local/share/speedypy/speedtests.log.
(A nicer way to identify such servers will be included in future versions.)
Because we can exclude bad servers and the plot is done using a rolling average over many data points per day, the presented data should be quite robust and suitable for judging whether your actual bandwidth matches what you are paying for.
However, note that the connection speed may also be influenced by factors between your computer and your provider, such as a bad cable, faulty hardware or a poor wifi signal. Best results are achieved when running this on a computer that is connected via ethernet cable directly to the modem or router of your internet service provider.
Of course, this tool may not detect anomalies that occur in-between measurements, such as a complete outage of the connection with a duration of half an hour or less.
Adrian Vollmer
MIT