Motivation

Developers tend to misinterpret error handling in workflows with infrastructure errors (e.g. Server Error 503, Not Found 404, and so on). We need to explain how to properly do infrastructure error handling outside of workflows.

Remote calls can fail, or hang without a response for the length of timeouts (e.g. TCP). In such cases many workflows on a unresponsive supplier can lead to running out of critical resources leading to cascading failures across multiple systems. Circuit Breaker pattern prevents this kind of catastrophic cascade so that the error is handled quickly and gracefully without waiting for the timeout.

Goal

Write a dedicated guide to explain infrastructure error handling outside of workflows

Scenarios

As a developer, I need to detect failures so that I can encapsulate the logic of preventing a failure from constantly recurring, during maintenance, temporary external system failure or unexpected system difficulties.

Expected outcomes

A guide on how to use circuit breakers with workflows to deal with infrastructure failures

Adding circuit breaker will negatively affect performance and consumes additional resources on the system