Remove replica selection change FF for LB

Removes the feature flag. As the result,
there would be a lower chance the retry would
be required.

Changelog: performance

config/feature_flags/development/sidekiq_load_balancing_rotate_up_to_date_replica.yml

----
-name: sidekiq_load_balancing_rotate_up_to_date_replica
-introduced_by_url: https://gitlab.com/gitlab-org/gitlab/-/merge_requests/63413/
-rollout_issue_url: https://gitlab.com/gitlab-org/gitlab/-/issues/333153
-milestone: '14.0'
-type: development
-group: group::memory
-default_enabled: false

lib/gitlab/database/load_balancing/sidekiq_server_middleware.rb

@@ -68,11 +68,7 @@ module Gitlab
         end
 
         def replica_caught_up?(location)
-          if Feature.enabled?(:sidekiq_load_balancing_rotate_up_to_date_replica)
-            load_balancer.select_up_to_date_host(location)
-          else
-            load_balancer.host.caught_up?(location)
-          end
+          load_balancer.select_up_to_date_host(location)
         end
       end
     end

spec/lib/gitlab/database/load_balancing/sidekiq_server_middleware_spec.rb

@@ -156,20 +156,6 @@ RSpec.describe Gitlab::Database::LoadBalancing::SidekiqServerMiddleware do
             expect(job['load_balancing_strategy']).to eq('retry_primary')
           end
         end
-
-        context 'replica selection mechanism feature flag rollout' do
-          before do
-            stub_feature_flags(sidekiq_load_balancing_rotate_up_to_date_replica: false)
-          end
-
-          it 'uses different implementation' do
-            expect(::Gitlab::Database::LoadBalancing).to receive_message_chain(:proxy, :load_balancer, :host, :caught_up?).and_return(false)
-
-            expect do
-              process_job(job)
-            end.to raise_error(Sidekiq::JobRetry::Skip)
-          end
-        end
       end
     end