EventFlow Go

Where work gets stuck

A missing integration update can leave an order in limbo, create support work and force staff to search logs across systems.

This project is for commerce platforms, saas teams and operations teams connecting separate services. The aim is to make the next decision visible to the people doing the work.

What a team might do

A retailer records order 1042. The fulfilment system responds with a temporary error, so the event remains visible and is tried again. An operator can inspect the attempts instead of guessing whether the order update disappeared. A repeated request with the same identity does not create a second event.

What the prototype does

• Authenticated event intake and stable idempotency keys
• PostgreSQL-backed attempts and retryable delivery
• Concurrent workers with leases and visible delivery history

What I built and tested

I designed and implemented this independent prototype to explore the business workflow and its technical constraints. Go, HTTP APIs, PostgreSQL transactions, worker concurrency, signed webhooks and tests. The repository explains how to run its real implementation and tests.

Explore the demo and the code

Open the fictional browser walkthrough to understand the workflow. Read the source and README for installation, local testing and implementation details. The browser page does not run the backend service.

What this version does not claim

Could this fit your team?

Choose one important business event and map what staff do when its destination is unavailable. Test the workflow using fictional events before discussing a real integration. There is no fee or obligation to adopt a product. Any work involving real data or a production deployment would need a separate scope and appropriate safeguards.

Tools and skills