DEV: Always enqueue sidekiq jobs after database transaction commit (#11293)

When jobs are enqueued inside a transaction, it's possible that they will be executed before the necessary data is available in the database. This commit ensures all jobs are enqueued in an ActiveRecord after_commit hook.

One potential downside here is if the job fails to enqueue, the transaction will no longer be aborted. However, the chance of that happening is reasonably low, and the impact is significantly lower than the current issue where jobs are scheduled before their data is ready.

app/jobs/base.rb

@@ -311,7 +311,7 @@ module Jobs
         hash['queue'] = queue
       end
 
-      klass.client_push(hash)
+      DB.after_commit { klass.client_push(hash) }
     else
       # Otherwise execute the job right away
       opts.delete(:delay_for)

spec/jobs/jobs_spec.rb

@@ -29,6 +29,28 @@ describe Jobs do
         end
       end
 
+      it "enqueues the job after the current transaction has committed" do
+        jobs = Jobs::ProcessPost.jobs
+        expect(jobs.length).to eq(0)
+
+        Jobs.enqueue(:process_post, post_id: 1)
+        expect(jobs.length).to eq(1)
+
+        ActiveRecord::Base.transaction do
+          Jobs.enqueue(:process_post, post_id: 1)
+          expect(jobs.length).to eq(1)
+        end
+        expect(jobs.length).to eq(2)
+
+        # Failed transation
+        ActiveRecord::Base.transaction do
+          Jobs.enqueue(:process_post, post_id: 1)
+          raise ActiveRecord::Rollback
+        end
+
+        expect(jobs.length).to eq(2) # No change
+      end
+
       it "does not pass current_site_id when 'all_sites' is present" do
         Sidekiq::Testing.fake! do
           jobs = Jobs::ProcessPost.jobs