Hacker News is a community-driven platform for sharing and discussing technology news, startups, and programming-related topics. Through user submissions and comments, Hacker News offers insights into emerging technology trends, industry developments, and entrepreneurial ventures. Readers can participate in discussions, share their insights, and stay informed about the latest advancements in technology and innovation.

Hacker News

A major AWS outage in the US-EAST-1 region on October 20, 2025, caused by DNS resolution issues with DynamoDB endpoints, took 75 minutes to diagnose and affected much of the internet. The incident highlights concerns about AWS's loss of institutional knowledge due to significant employee departures, layoffs (27,000+ since 2022), and high regretted attrition rates (69-81%). Senior engineers who understood deep system failure modes have left, potentially leaving newer teams without the tribal knowledge needed to quickly detect and resolve complex infrastructure issues. The outage suggests that cost-cutting measures and talent drain may be compromising AWS's operational resilience.

Today is when Amazon brain drain finally caught up with AWS

There’s so many lessons to be learnt here, unfortunately those that should learn them will never learn them because … ego’s
Even giants trip and fall, putting so much trust in cloud infrastructure will be the demise of many businesses. Yes, use them, but don’t trust them.
The marketing will have you believe they’re invincible gods but behind that facade is a slowly (perhaps not so slowly according to this article) dimishing expertise/competence level, and eventually the cracks will show, more “stupid” mistakes will happen.
Hopefully at the point of no return there’s enough of us around still to rekindle the self hosted solution. Yes it’s hard and requires discipline and skill to keep running but what’s wrong with that? At least you’re in control.