It all started with a desire: I wanted some basic analytics for this blog to know how it’s
doing, which articles are popular and which are the main referrers. But therein lied the
conundrum: I am a privacy zealot, very much against extensive online tracking (just look
at the browser extensions I use), so using a
third-party service that collected all my visitors’ data was against the question.

I knew there were open-source web analytics projects out there, like Piwik which has been
around for a long time, but never had the time or need to look into them seriously. Until
now.

Please note that this article is by no means an extensive list. There are a lot of other
options out there, just check out this big
list. Each software has its own pros
and cons, but I hope this article can help you choose and make the jump towards
self-hosted analytics, or at least to abandon Google Analytics in favor of an
open-source-based SAAS option.

I researched options for my specific use case:

1. Must be free software (free as in freedom, not as in free beer) and self-hosted.

2. Simple deployment, the fewer moving parts the better. SQLite if possible so no DB
   server will be needed. No Redis/Memcached if it can be avoided. Option to deploy without
   Docker would be nice.

3. Reasonable resource usage. My sites don’t usually get a lot of traffic, so I would like
   to reuse the 2 GB VPS that hosts this blog.

4. Technology that I can easily hack is a plus. I will favor Python projects, but YMMV.

5. Lightweight and not too intrusive. I basically just want to know which articles are
   seen when, plus referrers, no need for advanced analytics.

6. Actively maintained.

Home page

1. GPLv3. Paid SAAS cloud service with free self-hosting option.

2. Installation docs.

   • Requires a webserver (Apache, Nginx, etc.)
   • Requires MySQL.
   • Configuration of Matomo is all done in-app.

3. Matomo self-hosted requirements:

   • <100k page views/month: 1 server for both app and DB. 2 CPU, 2 GB RAM, 50GB SSD disk.
   • ~1M