ActiveAnalytics

Simple traffic analytics for the win of privacy.

• NO cookies
• NO JavaScript
• NO third parties
• NO bullshit

ActiveAnalytics is a Rails engine directly mountable in your Ruby on Rails application. It doesn't reveal anything about specific visitors. It cannot be blocked by adblockers or other privacy-protecting extensions (and doesn't need to).

ActiveAnalytics lets you know about:

• Sources: What are the pages and domains that bring some traffic.
• Page views: What are the pages that are the most viewed in your application.
• Next/previous page: What are the entry and exit pages for a given page of your application.

Installation

Add this line to your application's Gemfile:

gem 'active_analytics'

Then execute bundle and run the migration:

bundle
rails active_analytics:install:migrations
rails db:migrate

Add the route to ActiveAnalytics dashboard at the desired endpoint:

# config/routes.rb
mount ActiveAnalytics::Engine, at: "analytics"  # http://localhost:3000/analytics

By default ActiveAnalytics will extend ActionController::Base, but you can specify a custom base controller for the ActiveAnalytics dashboard:

# config/initializers/active_analytics.rb
Rails.application.configure do
  ActiveAnalytics.base_controller_class = "AdminController"
end

The next step is to collect trafic and there is 2 options.

Record requests synchronously

This is the easiest way to start with. However it's less performant since it triggers a write into your database for each request. Your controllers have to call ActiveAnalytics.record_request(request) to record page views. The Rails way to achieve is to use after_action :

class ApplicationController < ActionController::Base
  after_action :record_page_view

  def record_page_view
    # This is a basic example, you might need to customize some conditions.
    # For most sites, it makes no sense to record anything other than HTML.
    if response.content_type && response.content_type.start_with?("text/html")
      # Add a condition to record only your canonical domain
      # and use a gem such as crawler_detect to skip bots.
      ActiveAnalytics.record_request(request)
    end
  end
end

In case you don't want to record all page views, because each application has sensitive URLs such as password reset and so on, simply define a skip_after_action :record_page_view in the relevant controller.

Queue requests asynchronously

It requires more work and it's relevant if your application handle a large trafic. The idea is to queue data into Redis because it does not require the database writing to the disk on each request. First you have to set the Redis URL or connection.

# File lib/patches/active_analytics.rb or config/initializers/active_analytics.rb

ActiveAnalytics.redis_url = "redis://user:password@host/1" # Default ENV["ACTIVE_ANALYTICS_REDIS_URL"] || ENV["REDIS_URL"] || "redis://localhost"

# If you use special connection options you have to instantiate it yourself
ActiveAnalytics.redis = Redis.new(
  url: ENV["REDIS_URL"],
  reconnect_attempts: 10,
  ssl_params: {verify_mode: OpenSSL::SSL::VERIFY_NONE}
)

Then your controllers have to call ActiveAnalytics.queue_request(request) to queue page views. The Rails way to achieve is to use after_action :

class ApplicationController < ActionController::Base
  after_action :queue_page_view

  def queue_page_view
    # This is a basic example, you might need to customize some conditions.
    # For most sites, it makes no sense to record anything other than HTML.
    if response.content_type && response.content_type.start_with?("text/html")
      # Add a condition to record only your canonical domain
      # and use a gem such as crawler_detect to skip bots.
      ActiveAnalytics.queue_request(request)
    end
  end
end

Queued data need to be saved into the database in order to be viewable in the ActiveAnalytics dashboard. For that, call ActiveAnalytics.flush_queue from a cron task or a background job.

It's up to you if you want to flush the queue every hour or every 10 minutes. I advise to execute the last flush of the day at 23:59. It prevents from shifting the trafic to the next day. In that case only the last minute will be shifted to the next day, even if the flush ends after midnight. This small imperfection allows a simpler implementation for now. Keep it simple !

Authentication and permissions

ActiveAnalytics cannot guess how you handle user authentication, because it is different for all Rails applications. So you have to monkey patch ActiveAnalytics::ApplicationController in order to inject your own mechanism. The patch can be saved wherever you want. For example, I like to have all the patches in one place, so I put them in lib/patches.

# lib/patches/active_analytics.rb

ActiveAnalytics::ApplicationController.class_eval do
    before_action :require_admin

    def require_admin
      # This example supposes there are current_user and User#admin? methods
      raise ActionController::RoutingError.new("Not found") unless current_user.try(:admin?)
    end
  end
end

Then you have to require the monkey patch. Because it's loaded via require, it won't be reloaded in development. Since you are not supposed to change this file often, it should not be an issue.

# config/application.rb
config.after_initialize do
  require "patches/active_analytics"
end

If you use Devise, you can check the permission directly from routes.rb :

# config/routes.rb
authenticate :user, -> (u) { u.admin? } do # Supposing there is a User#admin? method
  mount ActiveAnalytics::Engine, at: "analytics" # http://localhost:3000/analytics
end

License

The gem is available as open-source under the terms of the MIT License.

Made by Base Secrète.

Rails developer? Check out RoRvsWild, our Ruby on Rails application monitoring tool.