Migrates Snowplow backend from EE to CE

This introduces several changes, but these are all just ported from the
EE project.

Gemfile

@@ -297,6 +297,9 @@ gem 'batch-loader', '~> 1.4.0'
 # Perf bar
 gem 'peek', '~> 1.0.1'
+# Snowplow events tracking
+gem 'snowplow-tracker', '~> 0.6.1'
+
 # Memory benchmarks
 gem 'derailed_benchmarks', require: false

Gemfile.lock

@@ -152,6 +152,7 @@ GEM
     concurrent-ruby-ext (1.1.5)
       concurrent-ruby (= 1.1.5)
     connection_pool (2.2.2)
+    contracts (0.11.0)
     crack (0.4.3)
       safe_yaml (~> 1.0.0)
     crass (1.0.4)
@@ -901,6 +902,8 @@ GEM
       simplecov-html (~> 0.10.0)
     simplecov-html (0.10.2)
     slack-notifier (1.5.1)
+    snowplow-tracker (0.6.1)
+      contracts (~> 0.7, <= 0.11)
     spring (2.0.2)
       activesupport (>= 4.2)
     spring-commands-rspec (1.0.4)
@@ -1229,6 +1232,7 @@ DEPENDENCIES
   simple_po_parser (~> 1.1.2)
   simplecov (~> 0.16.1)
   slack-notifier (~> 1.5.1)
+  snowplow-tracker (~> 0.6.1)
   spring (~> 2.0.0)
   spring-commands-rspec (~> 1.0.4)
   sprockets (~> 3.7.0)

app/helpers/application_settings_helper.rb

@@ -270,7 +270,11 @@ module ApplicationSettingsHelper
       :diff_max_patch_bytes,
       :commit_email_hostname,
       :protected_ci_variables,
-      :local_markdown_version
+      :local_markdown_version,
+      :snowplow_collector_hostname,
+      :snowplow_cookie_domain,
+      :snowplow_enabled,
+      :snowplow_site_id
     ]
   end

app/helpers/tracking_helper.rb

@@ -2,6 +2,21 @@
 module TrackingHelper
   def tracking_attrs(label, event, property)
-    {} # CE has no tracking features
+    return {} unless tracking_enabled?
+
+    {
+      data: {
+        track_label: label,
+        track_event: event,
+        track_property: property
+      }
+    }
+  end
+
+  private
+
+  def tracking_enabled?
+    Rails.env.production? &&
+      ::Gitlab::CurrentSettings.snowplow_enabled?
   end
 end

app/models/application_setting.rb

@@ -99,6 +99,11 @@ class ApplicationSetting < ApplicationRecord
             presence: true,
             if: :plantuml_enabled
+  validates :snowplow_collector_hostname,
+            presence: true,
+            hostname: true,
+            if: :snowplow_enabled
+
   validates :max_attachment_size,
             presence: true,
             numericality: { only_integer: true, greater_than: 0 }

app/models/application_setting_implementation.rb

@@ -97,6 +97,10 @@ module ApplicationSettingImplementation
         usage_stats_set_by_user_id: nil,
         diff_max_patch_bytes: Gitlab::Git::Diff::DEFAULT_MAX_PATCH_BYTES,
         commit_email_hostname: default_commit_email_hostname,
+        snowplow_collector_hostname: nil,
+        snowplow_cookie_domain: nil,
+        snowplow_enabled: false,
+        snowplow_site_id: nil,
         protected_ci_variables: false,
         local_markdown_version: 0,
         outbound_local_requests_whitelist: [],

app/views/admin/application_settings/_snowplow.html.haml

+- expanded = true if !@application_setting.valid? && @application_setting.errors.any? { |k| k.to_s.start_with?('snowplow_') }
+%section.settings.as-snowplow.no-animate#js-snowplow-settings{ class: ('expanded' if expanded) }
+  .settings-header
+    %h4
+      = _('Snowplow')
+    %button.btn.btn-default.js-settings-toggle{ type: 'button' }
+      = expanded ? _('Collapse') : _('Expand')
+    %p
+      = _('Configure the %{link} integration.').html_safe % { link: link_to('Snowplow', 'https://snowplowanalytics.com/', target: '_blank') }
+  .settings-content
+
+    = form_for @application_setting, url: integrations_admin_application_settings_path, html: { class: 'fieldset-form' } do |f|
+      = form_errors(@application_setting)
+
+      %fieldset
+        .form-group
+          .form-check
+            = f.check_box :snowplow_enabled, class: 'form-check-input'
+            = f.label :snowplow_enabled, _('Enable snowplow tracking'), class: 'form-check-label'
+        .form-group
+          = f.label :snowplow_collector_hostname, _('Collector hostname'), class: 'label-light'
+          = f.text_field :snowplow_collector_hostname, class: 'form-control', placeholder: 'snowplow.example.com'
+        .form-group
+          = f.label :snowplow_site_id, _('Site ID'), class: 'label-light'
+          = f.text_field :snowplow_site_id, class: 'form-control'
+        .form-group
+          = f.label :snowplow_cookie_domain, _('Cookie domain'), class: 'label-light'
+          = f.text_field :snowplow_cookie_domain, class: 'form-control'
+
+      = f.submit _('Save changes'), class: 'btn btn-success'

app/views/layouts/_snowplow.html.haml

+- return unless Gitlab::CurrentSettings.snowplow_enabled?
+
+= javascript_tag nonce: true do
+  :plain
+    ;(function(p,l,o,w,i,n,g){if(!p[i]){p.GlobalSnowplowNamespace=p.GlobalSnowplowNamespace||[];
+    p.GlobalSnowplowNamespace.push(i);p[i]=function(){(p[i].q=p[i].q||[]).push(arguments)
+    };p[i].q=p[i].q||[];n=l.createElement(o);g=l.getElementsByTagName(o)[0];n.async=1;
+    n.src=w;g.parentNode.insertBefore(n,g)}}(window,document,"script","#{asset_url('snowplow/sp.js')}","snowplow"));
+
+    window.snowplow('newTracker', '#{Gitlab::SnowplowTracker::NAMESPACE}', '#{Gitlab::CurrentSettings.snowplow_collector_hostname}', {
+      appId: '#{Gitlab::CurrentSettings.snowplow_site_id}',
+      cookieDomain: '#{Gitlab::CurrentSettings.snowplow_cookie_domain}',
+      userFingerprint: false,
+      respectDoNotTrack: true,
+      forceSecureTracker: true,
+      post: true,
+      contexts: { webPage: true },
+      stateStorageStrategy: "localStorage"
+    });
+
+    window.snowplow('enableActivityTracking', 30, 30);
+    window.snowplow('trackPageView');
+
+- return unless Feature.enabled?(:additional_snowplow_tracking, @group)
+
+= javascript_tag nonce: true do
+  :plain
+    window.snowplow('enableFormTracking');
+    window.snowplow('enableLinkClickTracking');

doc/api/settings.md

@@ -321,4 +321,8 @@ are listed in the descriptions of the relevant settings.
 | `user_show_add_ssh_key_message`          | boolean          | no                                   | When set to `false` disable the "You won't be able to pull or push project code via SSH" warning shown to users with no uploaded SSH key. |
 | `version_check_enabled`                  | boolean          | no                                   | Let GitLab inform you when an update is available. |
 | `local_markdown_version`                 | integer          | no                                   | Increase this value when any cached markdown should be invalidated. |
+| `snowplow_enabled`                       | boolean          | no                                   | Enable snowplow tracking. |
+| `snowplow_collector_hostname`            | string           | required by: `snowplow_enabled`      | The Snowplow collector hostname. (e.g. `snowplow.trx.gitlab.net`) | 
+| `snowplow_site_id`                       | string           | no                                   | The Snowplow site name / application id. (e.g. `gitlab`) |
+| `snowplow_cookie_domain`                 | string           | no                                   | The Snowplow cookie domain. (e.g. `.gitlab.com`) |
 | `geo_node_allowed_ips`                   | string           | yes                                  | **(PREMIUM)** Comma-separated list of IPs and CIDRs of allowed secondary nodes. For example, `1.1.1.1, 2.2.2.0/24`. |

lib/api/settings.rb

@@ -125,6 +125,12 @@ module API
       optional :instance_statistics_visibility_private, type: Boolean, desc: 'When set to `true` Instance statistics will only be available to admins'
       optional :local_markdown_version, type: Integer, desc: "Local markdown version, increase this value when any cached markdown should be invalidated"
       optional :allow_local_requests_from_hooks_and_services, type: Boolean, desc: 'Deprecated: Use :allow_local_requests_from_web_hooks_and_services instead. Allow requests to the local network from hooks and services.' # support legacy names, can be removed in v5
+      optional :snowplow_enabled, type: Grape::API::Boolean, desc: 'Enable Snowplow tracking'
+      given snowplow_enabled: ->(val) { val } do
+        requires :snowplow_collector_hostname, type: String, desc: 'The Snowplow collector hostname'
+        optional :snowplow_cookie_domain, type: String, desc: 'The Snowplow cookie domain'
+        optional :snowplow_site_id, type: String, desc: 'The Snowplow site name / application ic'
+      end
       ApplicationSetting::SUPPORTED_KEY_TYPES.each do |type|
         optional :"#{type}_key_restriction",

lib/gitlab/snowplow_tracker.rb

+# frozen_string_literal: true
+
+require 'snowplow-tracker'
+
+module Gitlab
+  module SnowplowTracker
+    NAMESPACE = 'cf'
+
+    class << self
+      def track_event(category, action, label: nil, property: nil, value: nil, context: nil)
+        tracker&.track_struct_event(category, action, label, property, value, context, Time.now.to_i)
+      end
+
+      private
+
+      def tracker
+        return unless enabled?
+
+        @tracker ||= ::SnowplowTracker::Tracker.new(emitter, subject, NAMESPACE, Gitlab::CurrentSettings.snowplow_site_id)
+      end
+
+      def subject
+        ::SnowplowTracker::Subject.new
+      end
+
+      def emitter
+        ::SnowplowTracker::Emitter.new(Gitlab::CurrentSettings.snowplow_collector_hostname)
+      end
+
+      def enabled?
+        Gitlab::CurrentSettings.snowplow_enabled?
+      end
+    end
+  end
+end

locale/gitlab.pot

@@ -2953,6 +2953,9 @@ msgstr ""
 msgid "Collapse sidebar"
 msgstr ""
+msgid "Collector hostname"
+msgstr ""
+
 msgid "ComboSearch is not defined"
 msgstr ""
@@ -3120,6 +3123,9 @@ msgstr ""
 msgid "Configure storage path settings."
 msgstr ""
+msgid "Configure the %{link} integration."
+msgstr ""
+
 msgid "Configure the way a user creates a new account."
 msgstr ""
@@ -3261,6 +3267,9 @@ msgstr ""
 msgid "ConvDev Index"
 msgstr ""
+msgid "Cookie domain"
+msgstr ""
+
 msgid "Copied"
 msgstr ""
@@ -4253,6 +4262,9 @@ msgstr ""
 msgid "Enable shared Runners"
 msgstr ""
+msgid "Enable snowplow tracking"
+msgstr ""
+
 msgid "Enable two-factor authentication"
 msgstr ""
@@ -10286,6 +10298,9 @@ msgstr ""
 msgid "Similar issues"
 msgstr ""
+msgid "Site ID"
+msgstr ""
+
 msgid "Size and domain settings for static websites"
 msgstr ""
@@ -10316,6 +10331,9 @@ msgstr ""
 msgid "SnippetsEmptyState|They can be either public or private."
 msgstr ""
+msgid "Snowplow"
+msgstr ""
+
 msgid "Some email servers do not support overriding the email sender name. Enable this option to include the name of the author of the issue, merge request or comment in the email body instead."
 msgstr ""

spec/helpers/tracking_helper_spec.rb

@@ -4,8 +4,32 @@ require 'spec_helper'
 describe TrackingHelper do
   describe '#tracking_attrs' do
-    it 'returns an empty hash' do
-      expect(helper.tracking_attrs('a', 'b', 'c')).to eq({})
+    using RSpec::Parameterized::TableSyntax
+
+    let(:input) { %w(a b c) }
+    let(:results) do
+      {
+        no_data: {},
+        with_data: { data: { track_label: 'a', track_event: 'b', track_property: 'c' } }
+      }
+    end
+
+    where(:snowplow_enabled, :environment, :result) do
+      true  | 'production'  | :with_data
+      false | 'production'  | :no_data
+      true  | 'development' | :no_data
+      false | 'development' | :no_data
+      true  | 'test'        | :no_data
+      false | 'test'        | :no_data
+    end
+
+    with_them do
+      it 'returns a hash' do
+        stub_application_setting(snowplow_enabled: snowplow_enabled)
+        allow(Rails).to receive(:env).and_return(environment.inquiry)
+
+        expect(helper.tracking_attrs(*input)).to eq(results[result])
+      end
     end
   end
 end

spec/lib/gitlab/snowplow_tracker_spec.rb

+# frozen_string_literal: true
+require 'spec_helper'
+
+describe Gitlab::SnowplowTracker do
+  let(:timestamp) { Time.utc(2017, 3, 22) }
+
+  around do |example|
+    Timecop.freeze(timestamp) { example.run }
+  end
+
+  subject { described_class.track_event('epics', 'action', property: 'what', value: 'doit') }
+
+  context '.track_event' do
+    context 'when Snowplow tracker is disabled' do
+      it 'does not track the event' do
+        expect(SnowplowTracker::Tracker).not_to receive(:new)
+
+        subject
+      end
+    end
+
+    context 'when Snowplow tracker is enabled' do
+      before do
+        stub_application_setting(snowplow_enabled: true)
+        stub_application_setting(snowplow_site_id: 'awesome gitlab')
+        stub_application_setting(snowplow_collector_hostname: 'url.com')
+      end
+
+      it 'tracks the event' do
+        tracker = double
+
+        expect(::SnowplowTracker::Tracker).to receive(:new)
+          .with(
+            an_instance_of(::SnowplowTracker::Emitter),
+            an_instance_of(::SnowplowTracker::Subject),
+            'cf', 'awesome gitlab'
+          ).and_return(tracker)
+        expect(tracker).to receive(:track_struct_event)
+          .with('epics', 'action', nil, 'what', 'doit', nil, timestamp.to_i)
+
+        subject
+      end
+    end
+  end
+end

spec/requests/api/settings_spec.rb

@@ -144,6 +144,7 @@ describe API::Settings, 'Settings' do
           external_auth_client_key_pass: "5iveL!fe"
         }
       end
+
       let(:attribute_names) { settings.keys.map(&:to_s) }
       it 'includes the attributes in the API' do
@@ -165,6 +166,56 @@ describe API::Settings, 'Settings' do
       end
     end
+    context "snowplow tracking settings" do
+      let(:settings) do
+        {
+          snowplow_collector_hostname: "snowplow.example.com",
+          snowplow_cookie_domain: ".example.com",
+          snowplow_enabled: true,
+          snowplow_site_id: "site_id"
+        }
+      end
+
+      let(:attribute_names) { settings.keys.map(&:to_s) }
+
+      it "includes the attributes in the API" do
+        get api("/application/settings", admin)
+
+        expect(response).to have_gitlab_http_status(200)
+        attribute_names.each do |attribute|
+          expect(json_response.keys).to include(attribute)
+        end
+      end
+
+      it "allows updating the settings" do
+        put api("/application/settings", admin), params: settings
+
+        expect(response).to have_gitlab_http_status(200)
+        settings.each do |attribute, value|
+          expect(ApplicationSetting.current.public_send(attribute)).to eq(value)
+        end
+      end
+
+      context "missing snowplow_collector_hostname value when snowplow_enabled is true" do
+        it "returns a blank parameter error message" do
+          put api("/application/settings", admin), params: { snowplow_enabled: true }
+
+          expect(response).to have_gitlab_http_status(400)
+          expect(json_response["error"]).to eq("snowplow_collector_hostname is missing")
+        end
+
+        it "handles validation errors" do
+          put api("/application/settings", admin), params: settings.merge({
+            snowplow_collector_hostname: nil
+          })
+
+          expect(response).to have_gitlab_http_status(400)
+          message = json_response["message"]
+          expect(message["snowplow_collector_hostname"]).to include("can't be blank")
+        end
+      end
+    end
+
     context "missing plantuml_url value when plantuml_enabled is true" do
       it "returns a blank parameter error message" do
         put api("/application/settings", admin), params: { plantuml_enabled: true }

spec/views/layouts/_head.html.haml_spec.rb

@@ -70,6 +70,23 @@ describe 'layouts/_head' do
     expect(rendered).to match('<link rel="stylesheet" media="all" href="/stylesheets/highlight/themes/solarised-light.css" />')
   end
+  context 'when an asset_host is set and snowplow url is set' do
+    let(:asset_host) { 'http://test.host' }
+
+    before do
+      allow(ActionController::Base).to receive(:asset_host).and_return(asset_host)
+      allow(Gitlab::CurrentSettings).to receive(:snowplow_enabled?).and_return(true)
+      allow(Gitlab::CurrentSettings).to receive(:snowplow_collector_hostname).and_return('www.snow.plow')
+    end
+
+    it 'add a snowplow script tag with asset host' do
+      render
+      expect(rendered).to match('http://test.host/assets/snowplow/')
+      expect(rendered).to match('window.snowplow')
+      expect(rendered).to match('www.snow.plow')
+    end
+  end
+
   def stub_helper_with_safe_string(method)
     allow_any_instance_of(PageLayoutHelper).to receive(method)
       .and_return(%q{foo" http-equiv="refresh}.html_safe)