Update code stage calculation in VSA

This change updates the calculation in value stream analytics. For the
code stage, the start event timestamp can be first_commit_at (new) or
first_mentioned_in_commit_at.

The change also introduces a separate stage where only the
first_commit_at is used.

ee/changelogs/unreleased/add-first-commit-at-to-code-stage.yml

+---
+title: Include first commit date in code stage calculation in Value Stream Analytics
+merge_request: 47215
+author:
+type: changed

ee/lib/ee/gitlab/analytics/cycle_analytics/stage_events.rb

@@ -22,7 +22,8 @@ module EE
             ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestClosed => 105,
             ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLastEdited => 106,
             ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelAdded => 107,
-            ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelRemoved => 108
+            ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelRemoved => 108,
+            ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestFirstCommitAt => 109
           }.freeze
 
           EE_EVENTS = EE_ENUM_MAPPING.keys.freeze
@@ -74,6 +75,15 @@ module EE
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::IssueLabelAdded,
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::IssueLabelRemoved
             ],
+            ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestFirstCommitAt => [
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestClosed,
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestFirstDeployedToProduction,
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLastBuildStarted,
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLastBuildFinished,
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLastEdited,
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelAdded,
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelRemoved
+            ],
             ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestCreated => [
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestClosed,
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestFirstDeployedToProduction,
@@ -87,12 +97,14 @@ module EE
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestFirstDeployedToProduction,
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLastEdited,
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelAdded,
-              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelRemoved
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelRemoved,
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestFirstCommitAt
             ],
             ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestFirstDeployedToProduction => [
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLastEdited,
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelAdded,
-              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelRemoved
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelRemoved,
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestFirstCommitAt
             ],
             ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLastBuildStarted => [
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestClosed,
@@ -100,7 +112,8 @@ module EE
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLastEdited,
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestMerged,
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelAdded,
-              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelRemoved
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelRemoved,
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestFirstCommitAt
             ],
             ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLastBuildFinished => [
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestClosed,
@@ -108,14 +121,16 @@ module EE
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLastEdited,
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestMerged,
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelAdded,
-              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelRemoved
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelRemoved,
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestFirstCommitAt
             ],
             ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestMerged => [
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestClosed,
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestFirstDeployedToProduction,
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLastEdited,
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelAdded,
-              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelRemoved
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelRemoved,
+              ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestFirstCommitAt
             ],
             ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelAdded => [
               ::Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestLabelAdded,

ee/lib/gitlab/analytics/cycle_analytics/stage_events/merge_request_first_commit_at.rb

+# frozen_string_literal: true
+
+module Gitlab
+  module Analytics
+    module CycleAnalytics
+      module StageEvents
+        class MergeRequestFirstCommitAt < MetricsBasedStageEvent
+          def self.name
+            s_("CycleAnalyticsEvent|Merge request first commit time")
+          end
+
+          def self.identifier
+            :merge_request_first_commit_at
+          end
+
+          def object_type
+            MergeRequest
+          end
+
+          def timestamp_projection
+            mr_metrics_table[:first_commit_at]
+          end
+        end
+      end
+    end
+  end
+end

ee/spec/lib/gitlab/analytics/cycle_analytics/data_collector_spec.rb

@@ -288,6 +288,23 @@ RSpec.describe Gitlab::Analytics::CycleAnalytics::DataCollector do
         it_behaves_like 'custom cycle analytics stage'
       end
 
+      context 'between first commit at and merge request merged time' do
+        let(:start_event_identifier) { :merge_request_first_commit_at }
+        let(:end_event_identifier) { :merge_request_merged }
+
+        def create_data_for_start_event(example_class)
+          create(:merge_request, :merged, source_project: example_class.project).tap do |mr|
+            mr.metrics.update!(first_commit_at: Time.now)
+          end
+        end
+
+        def create_data_for_end_event(mr, example_class)
+          mr.metrics.update!(merged_at: Time.now)
+        end
+
+        it_behaves_like 'custom cycle analytics stage'
+      end
+
       context 'between merge request build started time and build finished time' do
         let(:start_event_identifier) { :merge_request_last_build_started }
         let(:end_event_identifier) { :merge_request_last_build_finished }
@@ -367,39 +384,83 @@ RSpec.describe Gitlab::Analytics::CycleAnalytics::DataCollector do
         it_behaves_like 'custom cycle analytics stage'
       end
 
-      context 'between code stage start time and merge request created time with label filter' do
+      context 'between code stage start time and merge request created time' do
         let(:start_event_identifier) { :code_stage_start }
         let(:end_event_identifier) { :merge_request_created }
 
-        before do
-          params[:label_name] = [label.name, other_label.name]
-        end
+        context 'when issue is referenced in the commit message' do
+          def create_data_for_start_event(example_class)
+            issue = create(:issue, project: example_class.project)
+
+            mr = create(:merge_request, {
+              source_project: example_class.project,
+              target_branch: example_class.project.default_branch,
+              description: "Description\n\nclosing #{issue.to_reference}",
+              allow_broken: true
+            })
+
+            MergeRequests::UpdateService.new(
+              example_class.project,
+              user,
+              assignees: [user]
+            ).execute(mr)
+
+            mr.metrics.update!(first_commit_at: Time.zone.now)
+            mr
+          end
 
-        def create_data_for_start_event(example_class)
-          issue = create(:issue, project: example_class.project)
-          issue.metrics.update!(first_mentioned_in_commit_at: Time.zone.now)
+          def create_data_for_end_event(mr, example_class)
+            mr.update!(created_at: Time.zone.now)
+          end
 
-          mr = create(:merge_request, {
-            source_project: example_class.project,
-            target_branch: example_class.project.default_branch,
-            description: "Description\n\nclosing #{issue.to_reference}",
-            allow_broken: true
-          })
+          it_behaves_like 'custom cycle analytics stage'
+        end
 
-          MergeRequests::UpdateService.new(
-            example_class.project,
-            user,
-            label_ids: [label.id, other_label.id]
-          ).execute(mr)
+        context 'when `first_commit_at` is present' do
+          def create_data_for_start_event(example_class)
+            mr = create(:merge_request, { source_project: example_class.project, target_branch: example_class.project.default_branch, allow_broken: true })
+            mr.metrics.update!(first_commit_at: Time.zone.now)
+            mr
+          end
 
-          mr
-        end
+          def create_data_for_end_event(mr, example_class)
+            mr.update!(created_at: Time.zone.now)
+          end
 
-        def create_data_for_end_event(mr, example_class)
-          mr.update!(created_at: Time.zone.now)
+          it_behaves_like 'custom cycle analytics stage'
         end
 
-        it_behaves_like 'custom cycle analytics stage'
+        context 'label filter' do
+          before do
+            params[:label_name] = [label.name, other_label.name]
+          end
+
+          def create_data_for_start_event(example_class)
+            issue = create(:issue, project: example_class.project)
+            issue.metrics.update!(first_mentioned_in_commit_at: Time.zone.now)
+
+            mr = create(:merge_request, {
+              source_project: example_class.project,
+              target_branch: example_class.project.default_branch,
+              description: "Description\n\nclosing #{issue.to_reference}",
+              allow_broken: true
+            })
+
+            MergeRequests::UpdateService.new(
+              example_class.project,
+              user,
+              label_ids: [label.id, other_label.id]
+            ).execute(mr)
+
+            mr
+          end
+
+          def create_data_for_end_event(mr, example_class)
+            mr.update!(created_at: Time.zone.now)
+          end
+
+          it_behaves_like 'custom cycle analytics stage'
+        end
       end
     end
   end

ee/spec/lib/gitlab/analytics/cycle_analytics/stage_events/merge_request_first_commit_at_spec.rb

+# frozen_string_literal: true
+
+require 'spec_helper'
+
+RSpec.describe Gitlab::Analytics::CycleAnalytics::StageEvents::MergeRequestFirstCommitAt do
+  it_behaves_like 'cycle analytics event'
+end

lib/gitlab/analytics/cycle_analytics/stage_events/code_stage_start.rb

@@ -18,24 +18,46 @@ module Gitlab
           end
 
           def timestamp_projection
-            issue_metrics_table[:first_mentioned_in_commit_at]
+            Arel::Nodes::NamedFunction.new('COALESCE', column_list)
           end
 
           override :column_list
           def column_list
-            [timestamp_projection]
+            [
+              issue_metrics_table[:first_mentioned_in_commit_at],
+              mr_metrics_table[:first_commit_at]
+            ]
           end
 
           # rubocop: disable CodeReuse/ActiveRecord
           def apply_query_customization(query)
-            issue_metrics_join = mr_closing_issues_table
-              .join(issue_metrics_table)
+            query
+              .joins(merge_requests_closing_issues_join)
+              .joins(issue_metrics_join)
+              .joins(mr_metrics_join)
+          end
+          # rubocop: enable CodeReuse/ActiveRecord
+
+          def issue_metrics_join
+            mr_closing_issues_table
+              .join(issue_metrics_table, Arel::Nodes::OuterJoin)
               .on(mr_closing_issues_table[:issue_id].eq(issue_metrics_table[:issue_id]))
               .join_sources
+          end
 
-            query.joins(:merge_requests_closing_issues).joins(issue_metrics_join)
+          def merge_requests_closing_issues_join
+            mr_table
+              .join(mr_closing_issues_table, Arel::Nodes::OuterJoin)
+              .on(mr_table[:id].eq(mr_closing_issues_table[:merge_request_id]))
+              .join_sources
+          end
+
+          def mr_metrics_join
+            mr_metrics_table
+              .join(mr_metrics_table, Arel::Nodes::OuterJoin)
+              .on(mr_metrics_table[:merge_request_id].eq(mr_table[:id]))
+              .join_sources
           end
-          # rubocop: enable CodeReuse/ActiveRecord
         end
       end
     end

locale/gitlab.pot

@@ -8207,6 +8207,9 @@ msgstr ""
 msgid "CycleAnalyticsEvent|Merge request created"
 msgstr ""
 
+msgid "CycleAnalyticsEvent|Merge request first commit time"
+msgstr ""
+
 msgid "CycleAnalyticsEvent|Merge request first deployed to production"
 msgstr ""
 

spec/lib/gitlab/analytics/cycle_analytics/stage_events/code_stage_start_spec.rb

@@ -15,7 +15,7 @@ RSpec.describe Gitlab::Analytics::CycleAnalytics::StageEvents::CodeStageStart do
 
     other_merge_request = create(:merge_request, source_project: project, source_branch: 'a', target_branch: 'master')
 
-    records = subject.apply_query_customization(MergeRequest.all)
+    records = subject.apply_query_customization(MergeRequest.all).where('merge_requests_closing_issues.issue_id IS NOT NULL')
     expect(records).to eq([merge_request])
     expect(records).not_to include(other_merge_request)
   end