BulkImports: Reduce memory consumption when importing Epics

Avoid creating new pipeline object when the importing dataset has more
than one page. Instead, call the `BulkImports#run` recursively for each
page, while there's more page to be processed.

MR: https://gitlab.com/gitlab-org/gitlab/-/merge_requests/51595

ee/lib/ee/bulk_imports/groups/pipelines/epics_pipeline.rb

@@ -16,9 +16,9 @@ module EE
 
           loader EE::BulkImports::Groups::Loaders::EpicsLoader
 
-          after_run do |context|
+          def after_run(context)
             if context.entity.has_next_page?(:epics)
-              self.new.run(context)
+              run(context)
             end
           end
         end

ee/spec/lib/ee/bulk_imports/groups/pipelines/epics_pipeline_spec.rb

@@ -26,15 +26,13 @@ RSpec.describe EE::BulkImports::Groups::Pipelines::EpicsPipeline do
     subject { described_class.new }
 
     it 'imports group epics into destination group' do
-      page1 = extractor_data(has_next_page: true, cursor: 'nextPageCursor')
-      page2 = extractor_data(has_next_page: false)
+      first_page = extractor_data(has_next_page: true, cursor: 'nextPageCursor')
+      last_page = extractor_data(has_next_page: false)
 
       allow_next_instance_of(BulkImports::Common::Extractors::GraphqlExtractor) do |extractor|
-        if entity.has_next_page?(:epics)
-          allow(extractor).to receive(:extract).and_return(page2)
-        else
-          allow(extractor).to receive(:extract).and_return(page1)
-        end
+        allow(extractor)
+          .to receive(:extract)
+          .and_return(first_page, last_page)
       end
 
       expect { subject.run(context) }.to change(::Epic, :count).by(2)

lib/bulk_imports/pipeline/runner.rb

@@ -24,7 +24,7 @@ module BulkImports
           end
         end
 
-        after_run.call(context) if after_run.present?
+        after_run(context) if respond_to?(:after_run)
       rescue MarkedAsFailedError
         log_skip(context)
       end