Skip to content
All work
Case study 04Open source

Fluvio

A Go/PostgreSQL job queue with a dead-letter queue, sequencing and workflow support.

RoleAuthor
Year—
AreaOpen source
Stack
GoPostgreSQL

The brief

Fluvio came out of my work on BloomingBet: a job queue written in Go that keeps its state in PostgreSQL. I built it and open-sourced it.

What it does

Dead-letter queue. Jobs that keep failing are parked instead of retried forever.

Sequencing. Jobs that must run in order, do.

Workflows. Multi-step work can be described as one workflow.

Where it runs

I later extended Fluvio into the background worker for LoveWorld Lyrics v2, where it sends transactional email and runs async jobs, with an admin dashboard for monitoring them.

Fluvio Console. The dashboard is its own open-source project, fluvio_ui: a React admin app with live queue telemetry, job listing and filters, queue pause and resume, and a view of the worker fleet. It ships as a Docker image that reads its settings at start.

OpenHR. The OpenHR AWS stack runs the Fluvio dashboard as its own Fargate service on a separate host, locked down with an IP allow-list and basic auth.

View on GitHub

How it fits together

The system, in 3D

Services: Your service, Fluvio, Workers, Dead-letter queue, PostgreSQL. Connections: Your service to Fluvio; Fluvio to Workers; Fluvio to Dead-letter queue; Fluvio to PostgreSQL.
Your service
Fluvio
Workers
Dead-letter queue
PostgreSQL
Next project
05BloomingBetBackend engineer