Set next retry time when checksum calculation fails

ee/app/services/geo/repository_verification_primary_service.rb

 module Geo
   class RepositoryVerificationPrimaryService
+    include Delay
     include Gitlab::Geo::ProjectLogHelpers
 
     def initialize(project)
@@ -33,12 +34,29 @@ module Geo
     end
 
     def update_repository_state!(type, checksum: nil, failure: nil)
+      retry_at, retry_count =
+        if failure.present?
+          retry_count = repository_state.public_send("#{type}_retry_count").to_i + 1 # rubocop:disable GitlabSecurity/PublicSend
+          [next_retry_time(retry_count), retry_count]
+        end
+
       repository_state.update!(
         "#{type}_verification_checksum" => checksum,
-        "last_#{type}_verification_failure" => failure
+        "last_#{type}_verification_failure" => failure,
+        "#{type}_retry_at" => retry_at,
+        "#{type}_retry_count" => retry_count
       )
     end
 
+    # To prevent the retry time from storing invalid dates in the database,
+    # cap the max time to a week plus some random jitter value.
+    def next_retry_time(retry_count)
+      proposed_time = Time.now + delay(retry_count).seconds
+      max_future_time = Time.now + 7.days + delay(1).seconds
+
+      [proposed_time, max_future_time].min
+    end
+
     def repository_state
       @repository_state ||= project.repository_state || project.build_repository_state
     end

ee/spec/services/geo/repository_verification_primary_service_spec.rb

@@ -18,7 +18,11 @@ describe Geo::RepositoryVerificationPrimaryService do
         repository_verification_checksum: 'f123',
         last_repository_verification_failure: nil,
         wiki_verification_checksum: 'e321',
-        last_wiki_verification_failure: nil
+        last_wiki_verification_failure: nil,
+        repository_retry_at: nil,
+        repository_retry_count: nil,
+        wiki_retry_at: nil,
+        wiki_retry_count: nil
       )
     end
 
@@ -38,7 +42,11 @@ describe Geo::RepositoryVerificationPrimaryService do
         repository_verification_checksum: 'f123',
         last_repository_verification_failure: nil,
         wiki_verification_checksum: 'e321',
-        last_wiki_verification_failure: nil
+        last_wiki_verification_failure: nil,
+        repository_retry_at: nil,
+        repository_retry_count: nil,
+        wiki_retry_at: nil,
+        wiki_retry_count: nil
       )
     end
 
@@ -58,7 +66,11 @@ describe Geo::RepositoryVerificationPrimaryService do
         repository_verification_checksum: 'f123',
         last_repository_verification_failure: nil,
         wiki_verification_checksum: 'e321',
-        last_wiki_verification_failure: nil
+        last_wiki_verification_failure: nil,
+        repository_retry_at: nil,
+        repository_retry_count: nil,
+        wiki_retry_at: nil,
+        wiki_retry_count: nil
       )
     end
 
@@ -89,7 +101,11 @@ describe Geo::RepositoryVerificationPrimaryService do
         repository_verification_checksum: 'f123',
         last_repository_verification_failure: nil,
         wiki_verification_checksum: 'e321',
-        last_wiki_verification_failure: nil
+        last_wiki_verification_failure: nil,
+        repository_retry_at: nil,
+        repository_retry_count: nil,
+        wiki_retry_at: nil,
+        wiki_retry_count: nil
       )
     end
 
@@ -100,7 +116,11 @@ describe Geo::RepositoryVerificationPrimaryService do
         repository_verification_checksum: '0000000000000000000000000000000000000000',
         last_repository_verification_failure: nil,
         wiki_verification_checksum: '0000000000000000000000000000000000000000',
-        last_wiki_verification_failure: nil
+        last_wiki_verification_failure: nil,
+        repository_retry_at: nil,
+        repository_retry_count: nil,
+        wiki_retry_at: nil,
+        wiki_retry_count: nil
       )
     end
 
@@ -114,26 +134,59 @@ describe Geo::RepositoryVerificationPrimaryService do
         repository_verification_checksum: '0000000000000000000000000000000000000000',
         last_repository_verification_failure: nil,
         wiki_verification_checksum: '0000000000000000000000000000000000000000',
-        last_wiki_verification_failure: nil
+        last_wiki_verification_failure: nil,
+        repository_retry_at: nil,
+        repository_retry_count: nil,
+        wiki_retry_at: nil,
+        wiki_retry_count: nil
       )
     end
 
-    it 'keeps track of failures when calculating the repository checksum' do
+    context 'when checksum calculation fails' do
+      before do
         stub_project_repository(project, repository)
         stub_wiki_repository(project.wiki, wiki)
 
         allow(repository).to receive(:checksum).and_raise('Something went wrong with repository')
         allow(wiki).to receive(:checksum).twice.and_raise('Something went wrong with wiki')
+      end
 
+      it 'keeps track of failures' do
         subject.execute
 
         expect(project.repository_state).to have_attributes(
           repository_verification_checksum: nil,
           last_repository_verification_failure: 'Something went wrong with repository',
           wiki_verification_checksum: nil,
-        last_wiki_verification_failure: 'Something went wrong with wiki'
+          last_wiki_verification_failure: 'Something went wrong with wiki',
+          repository_retry_at: be_present,
+          repository_retry_count: 1,
+          wiki_retry_at: be_present,
+          wiki_retry_count: 1
         )
       end
+
+      it 'ensures the next retry time is capped properly' do
+        repository_state =
+          create(:repository_state,
+            project: project,
+            repository_retry_count: 30,
+            wiki_retry_count: 30)
+
+        subject.execute
+
+        expect(repository_state.reload).to have_attributes(
+          repository_verification_checksum: nil,
+          last_repository_verification_failure: 'Something went wrong with repository',
+          wiki_verification_checksum: nil,
+          last_wiki_verification_failure: 'Something went wrong with wiki',
+          repository_retry_at: be_within(100.seconds).of(Time.now + 7.days),
+          repository_retry_count: 31,
+          wiki_retry_at: be_within(100.seconds).of(Time.now + 7.days),
+          wiki_retry_count: 31
+        )
+      end
+    end
   end
 
   def stub_project_repository(project, repository)