Github Importer: Add Cache to Pull Request Reviews importer

Github importer have a mechanism to re-schedule an Importer job if it
gets throttled. But, repositories with large number of pull requests,
were in a _throttling loop_ for the Pull Request Reviews Importer.

In other words, when re-scheduled, the job was re-starting importing
reviews to all imported Pull Requests, regardless if already imported
or not, causing another throttling.

To fix this, already imported Pull Requests ids are being cached and
skipped, this way, when the job re-start, we don't do requests related
to already imported Pull Requests.

The same is applied to the Pull Request MergedBy importer.

Related to: https://gitlab.com/gitlab-org/gitlab/-/issues/329552

Changelog: changed
MR: https://gitlab.com/gitlab-org/gitlab/-/merge_requests/60668

changelogs/unreleased/kassio-github-importer-add-missing-importes-caches.yml

+---
+title: 'Github Importer: Add Cache to Pull Request Reviews importer'
+merge_request: 60668
+author:
+type: changed

lib/gitlab/github_import/client.rb

@@ -70,7 +70,7 @@ module Gitlab
       end
 
       def pull_request_reviews(repo_name, iid)
-        with_rate_limit { octokit.pull_request_reviews(repo_name, iid) }
+        each_object(:pull_request_reviews, repo_name, iid)
       end
 
       # Returns the details of a GitHub repository.

lib/gitlab/github_import/importer/pull_requests_merged_by_importer.rb

@@ -22,14 +22,18 @@ module Gitlab
           :pull_requests_merged_by
         end
 
-        def id_for_already_imported_cache(pr)
-          pr.number
+        def id_for_already_imported_cache(merge_request)
+          merge_request.id
         end
 
         def each_object_to_import
           project.merge_requests.with_state(:merged).find_each do |merge_request|
+            next if already_imported?(merge_request)
+
             pull_request = client.pull_request(project.import_source, merge_request.iid)
             yield(pull_request)
+
+            mark_as_imported(merge_request)
           end
         end
       end

lib/gitlab/github_import/importer/pull_requests_reviews_importer.rb

@@ -22,17 +22,22 @@ module Gitlab
           :pull_request_reviews
         end
 
-        def id_for_already_imported_cache(review)
-          review.github_id
+        def id_for_already_imported_cache(merge_request)
+          merge_request.id
         end
 
         def each_object_to_import
           project.merge_requests.find_each do |merge_request|
-            reviews = client.pull_request_reviews(project.import_source, merge_request.iid)
-            reviews.each do |review|
+            next if already_imported?(merge_request)
+
+            client
+              .pull_request_reviews(project.import_source, merge_request.iid)
+              .each do |review|
                 review.merge_request_id = merge_request.id
                 yield(review)
               end
+
+            mark_as_imported(merge_request)
           end
         end
       end

spec/lib/gitlab/github_import/client_spec.rb

@@ -32,8 +32,9 @@ RSpec.describe Gitlab::GithubImport::Client do
     it 'returns the pull request reviews' do
       client = described_class.new('foo')
 
-      expect(client.octokit).to receive(:pull_request_reviews).with('foo/bar', 999)
-      expect(client).to receive(:with_rate_limit).and_yield
+      expect(client)
+        .to receive(:each_object)
+        .with(:pull_request_reviews, 'foo/bar', 999)
 
       client.pull_request_reviews('foo/bar', 999)
     end

spec/lib/gitlab/github_import/importer/pull_requests_merged_by_importer_spec.rb

@@ -23,12 +23,11 @@ RSpec.describe Gitlab::GithubImport::Importer::PullRequestsMergedByImporter do
   end
 
   describe '#id_for_already_imported_cache' do
-    it { expect(subject.id_for_already_imported_cache(double(number: 1))).to eq(1) }
+    it { expect(subject.id_for_already_imported_cache(double(id: 1))).to eq(1) }
   end
 
-  describe '#each_object_to_import' do
+  describe '#each_object_to_import', :clean_gitlab_redis_cache do
     it 'fetchs the merged pull requests data' do
-      pull_request = double
       create(
         :merged_merge_request,
         iid: 999,
@@ -36,12 +35,18 @@ RSpec.describe Gitlab::GithubImport::Importer::PullRequestsMergedByImporter do
         target_project: project
       )
 
+      pull_request = double
+
       allow(client)
         .to receive(:pull_request)
+        .exactly(:once) # ensure to be cached on the second call
         .with('http://somegithub.com', 999)
         .and_return(pull_request)
 
-      expect { |b| subject.each_object_to_import(&b) }.to yield_with_args(pull_request)
+      expect { |b| subject.each_object_to_import(&b) }
+        .to yield_with_args(pull_request)
+
+      subject.each_object_to_import {}
     end
   end
 end

spec/lib/gitlab/github_import/importer/pull_requests_reviews_importer_spec.rb

@@ -23,12 +23,18 @@ RSpec.describe Gitlab::GithubImport::Importer::PullRequestsReviewsImporter do
   end
 
   describe '#id_for_already_imported_cache' do
-    it { expect(subject.id_for_already_imported_cache(double(github_id: 1))).to eq(1) }
+    it { expect(subject.id_for_already_imported_cache(double(id: 1))).to eq(1) }
   end
 
-  describe '#each_object_to_import' do
+  describe '#each_object_to_import', :clean_gitlab_redis_cache do
     it 'fetchs the merged pull requests data' do
-      merge_request = create(:merge_request, source_project: project)
+      merge_request = create(
+        :merged_merge_request,
+        iid: 999,
+        source_project: project,
+        target_project: project
+      )
+
       review = double
 
       expect(review)
@@ -37,10 +43,14 @@ RSpec.describe Gitlab::GithubImport::Importer::PullRequestsReviewsImporter do
 
       allow(client)
         .to receive(:pull_request_reviews)
+        .exactly(:once) # ensure to be cached on the second call
         .with('github/repo', merge_request.iid)
         .and_return([review])
 
-      expect { |b| subject.each_object_to_import(&b) }.to yield_with_args(review)
+      expect { |b| subject.each_object_to_import(&b) }
+        .to yield_with_args(review)
+
+      subject.each_object_to_import {}
     end
   end
 end