Merge branch 'optimize-running-stuck-5' into 'master'

Optimize StuckCiJobsWorker running builds query

See merge request gitlab-org/gitlab!71013

app/services/ci/stuck_builds/drop_running_service.rb

@@ -16,7 +16,12 @@ module Ci
       private
 
       def running_timed_out_builds
-        Ci::Build.running.updated_at_before(BUILD_RUNNING_OUTDATED_TIMEOUT.ago)
+        if Feature.enabled?(:ci_new_query_for_running_stuck_jobs, default_enabled: :yaml)
+          running_builds = Ci::Build.running.created_at_before(BUILD_RUNNING_OUTDATED_TIMEOUT.ago).order(created_at: :asc, project_id: :asc) # rubocop: disable CodeReuse/ActiveRecord
+          Ci::Build.id_in(running_builds).updated_at_before(BUILD_RUNNING_OUTDATED_TIMEOUT.ago)
+        else
+          Ci::Build.running.updated_at_before(BUILD_RUNNING_OUTDATED_TIMEOUT.ago)
+        end
       end
     end
   end

config/feature_flags/development/ci_new_query_for_running_stuck_jobs.yml

+---
+name: ci_new_query_for_running_stuck_jobs
+introduced_by_url: https://gitlab.com/gitlab-org/gitlab/-/merge_requests/71013
+rollout_issue_url: https://gitlab.com/gitlab-org/gitlab/-/issues/339264
+milestone: '14.4'
+type: development
+group: group::pipeline execution
+default_enabled: false

spec/services/ci/stuck_builds/drop_running_service_spec.rb

@@ -17,20 +17,47 @@ RSpec.describe Ci::StuckBuilds::DropRunningService do
     job.update!(job_attributes)
   end
 
-  context 'when job is running' do
-    let(:status) { 'running' }
+  around do |example|
+    freeze_time { example.run }
+  end
+
+  shared_examples 'running builds' do
+    context 'when job is running' do
+      let(:status) { 'running' }
+      let(:outdated_time) { described_class::BUILD_RUNNING_OUTDATED_TIMEOUT.ago - 30.minutes }
+      let(:fresh_time) { described_class::BUILD_RUNNING_OUTDATED_TIMEOUT.ago + 30.minutes }
+
+      context 'when job is outdated' do
+        let(:created_at) { outdated_time }
+        let(:updated_at) { outdated_time }
+
+        it_behaves_like 'job is dropped'
+      end
 
-    context 'when job was updated_at more than an hour ago' do
-      let(:updated_at) { 2.hours.ago }
+      context 'when job is fresh' do
+        let(:created_at) { fresh_time }
+        let(:updated_at) { fresh_time }
 
-      it_behaves_like 'job is dropped'
+        it_behaves_like 'job is unchanged'
+      end
+
+      context 'when job freshly updated' do
+        let(:created_at) { outdated_time }
+        let(:updated_at) { fresh_time }
+
+        it_behaves_like 'job is unchanged'
+      end
     end
+  end
 
-    context 'when job was updated in less than 1 hour ago' do
-      let(:updated_at) { 30.minutes.ago }
+  include_examples 'running builds'
 
-      it_behaves_like 'job is unchanged'
+  context 'when ci_new_query_for_running_stuck_jobs flag is disabled' do
+    before do
+      stub_feature_flags(ci_new_query_for_running_stuck_jobs: false)
     end
+
+    include_examples 'running builds'
   end
 
   %w(success skipped failed canceled scheduled pending).each do |status|
@@ -51,15 +78,4 @@ RSpec.describe Ci::StuckBuilds::DropRunningService do
       end
     end
   end
-
-  context 'for deleted project' do
-    let(:status) { 'running' }
-    let(:updated_at) { 2.days.ago }
-
-    before do
-      job.project.update!(pending_delete: true)
-    end
-
-    it_behaves_like 'job is dropped'
-  end
 end