Backpressure: what happens when your consumers can’t keep up

OccupySaaS

OccupySaaS is a protest movement in code: laid-off developers building clean-room, open-source, drop-in replacements for the SaaS giants. First up: ZeroAuthy (Auth0) and SalesCompel (Salesforce).

How Anthropic Redesigned Source Control

In this article, we’ll look at the architecture behind Scion.

Build Your Own Bank In Go

A streaming series is now available for free.

Apache Airflow 3.4.0: Task Loops, Dags in Any Language, and AI Agents

Apache Airflow 3.4.0 introduces bounded Task Loops for iterative workflows, Dag authoring beyond Python (TypeScript, Go, Java, and YAML) on the new AIP-85 Dag Importer, and a maturing Common AI provider for running LLM and agent tasks with built-in human approval.

Outbox Pattern in .NET With Postgres: Concurrent Workers vs the WAL

An API can save an order and crash before publishing the message that tells other services about it.

TinyGo 0.44 - Smaller, Faster, Better

Announcing TinyGo 0.44 with smaller and faster programs, more multicores on microcontrollers, support for Arduino UNO R4, new drivers, and big Bluetooth updates!

BoxBox: A Self-Hosted File Manager Without the Cloud Suite

BoxBox is a self-hosted file manager for browsing files already on a Linux server or NAS, without a full cloud suite like Nextcloud. It supports multiple storage locations in one interface, chunked uploads that resume after interruptions, and background jobs with progress updates for moving files. A built-in Monaco-based editor (the same tech used by VS Code) lets you edit configs in the browser. It also supports file and folder sharing with expiring links and ZIP downloads. BoxBox runs as a single Docker container with a Go backend and SvelteKit frontend, needs no separate database, and is released under the MIT license.

5 Event Driven Architecture Problems (And How to Avoid Them)

Event-driven architecture (EDA) phổ biến nhưng tiềm nhiều thách thức trong production environment. Vấn đề truy xuất lỗi (tracing) trở nên phức tạp khi sự kiện (event) đi qua nhiều service khác nhau, đòi hỏi hệ thống phân tích log tập trung như ELK stack. Hàng đợi tin hỏi (dead letter queues) dễ bị đầy nếu không implement validation logic kỹ càng, dẫn đến mất dữ kiện quan trọng. Sự không nhất quán cuối cùng (eventual consistency) gây race condition khi nhiều consumer xử lý cùng event đồng thời. Các EDA system cần thiết kế event schema versioning và comprehensive testing framework với Contract Testing để đảm bảo reliability khi số lượng service tăng lên.

Bài viết này giúp lập trình viên nhận diện và giải quyết các vấn đề phổ biến trong kiến trúc hướng sự kiện như tracing, dead letter queues, tính nhất quán cuối cùng và testing.