Matomo Webpage Analytics

So I am running this website for a while now but never cared if anyone else than people I tell about my website visit it. Yesterday, when talking to a friend, the question arised how many views this blog has.

As is have direct access to my /var/log/nginx/access.log, I can directly look into it. But it is much better having this done by something. That’s where Matomo comes into play.

Setting things up

I used the compose file and configuration from GitHub

version: "3"

services:
  db:
    image: mariadb:10
    command: --max-allowed-packet=64MB
    restart: always
    volumes:
      - ./data/db:/var/lib/mysql
    environment:
      - MYSQL_PASSWORD=matomo
      - MYSQL_DATABASE=matomo
      - MYSQL_USER=matomo
      - MARIADB_ROOT_PASSWORD=matomo

  app:
    image: matomo:fpm-alpine
    restart: always
    links:
      - db
    volumes:
      - ./data/matomo:/var/www/html
    environment:
      - MATOMO_DATABASE_HOST=db
      - PHP_MEMORY_LIMIT=2048M
      - MATOMO_DATABASE_ADAPTER=mysql
      - MATOMO_DATABASE_TABLES_PREFIX=matomo_
      - MATOMO_DATABASE_USERNAME=matomo
      - MATOMO_DATABASE_PASSWORD=matomo
      - MATOMO_DATABASE_DBNAME=matomo
  web:
    image: nginx:alpine
    restart: always
    volumes:
      - ./data/matomo:/var/www/html:ro
      # see https://github.com/matomo-org/matomo-nginx
      - ./matomo.conf:/etc/nginx/conf.d/default.conf:ro
    ports:
      - 8080:80

and started it without a lot of configuration. I am not having this open to the public, because I am not going to use a Cookie or JS-Script to analyze visitors behavior but just the page access to my webserver. Therefore I don’t need to change anything on my website.

I only need to pass the nginx-logs to matomo on a regular basis which I will describe shortly.

Sending logs

To do this, matomo allows to send logs to the running matomo service using the Matomo Server Log Analytics.

As I host multiple webpages on this server I am exactly using the vhosts nginx log format which has to be edited in /etc/nginx/nginx.conf

Finally one can import old logs using the following call to the script:

python3 data/matomo/misc/log-analytics/import_logs.py --log-format-name=common_complete --url 127.0.0.1:8080 --login matomo --password matomo --add-sites-new-hosts /var/log/nginx/access.log.1

But how will we do this regularly? The docs say that we need to add some log rotation. Fortunately, nginx has a default log-rotation installed

$ cat /etc/logrotate.d/nginx                                      
/var/log/nginx/*.log {
	daily
	missingok
	rotate 14
	compress
	delaycompress
	notifempty
	create 0640 www-data adm
	sharedscripts
	prerotate
		if [ -d /etc/logrotate.d/httpd-prerotate ]; then \
			run-parts /etc/logrotate.d/httpd-prerotate; \
		fi \
	endscript
	postrotate
		invoke-rc.d nginx rotate >/dev/null 2>&1
	endscript
}

which is run daily, and deletes logs older than 14 days, and leaves yesterdays logs without compression (delaycompress).

So we have a full day to import yesterdays log at /var/log/nginx/access.log.1 using a cron like this:

10 0 * * * python3 data/matomo/misc/log-analytics/import_logs.py --log-format-name=common_complete --url 127.0.0.1:8080 --login matomo --password matomo --add-sites-new-hosts /var/log/nginx/access.log.1

Aaand we are done and can see that our blog has only a handful of visits. But at least I have my projects documented and can forward it to people in need (mostly future me).