When working with a distributed microservice-based system, errors are bound to happen. These errors range from connectivity issues, serialization and deserialization problems, downstream system outages, and bugs, among others. To handle these issues, you need a simple and repeatable process. We can break down these errors into two categories: transient and non-transient errors. While transient errors can be fixed with retrying, you need human intervention to fix non-transient errors. In this post, we explore how to handle dead letter events and retry events for both types of errors while maintaining a flow of valid events. We also share some tips on how to notify someone and point them to the error event. By adopting this approach, you can increase the success of your project and minimize data loss.
