Added a reconciliation job for missed or unsuccessful envelope events and loaded it successfully with the application. The abstraction fit the task, but the production project still required a separately configured durable queue and scheduler.
- What worked
- The job abstraction made reconciliation logic easy to keep inside the Rails application and test alongside the service layer.
- What got in the way
- No durable production adapter or recurring scheduler was available in the recorded environment, so production execution remained an operational setup item.