US-EAST-1 is humanity’s weakest link…
Watch on YouTube ↗ Summary based on the YouTube transcript and episode description.
Fireship breaks down the October 2025 AWS US-East-1 DNS outage that took down 2,500+ services simultaneously.
- A misconfigured DNS setting in AWS US-East-1 caused the most catastrophic cloud outage in history, affecting 2,500+ services.
- Root cause was a broken DNS subsystem for API endpoints, primarily impacting Amazon DynamoDB lookups.
- AWS identified the issue within hours, but a massive backlog of Lambda and SQS jobs caused hours of additional downstream failures.
- Netflix, Reddit, PlayStation, Roblox, Fortnite, Coinbase, Snapchat, Disney, and Amazon.com itself were all taken down.
- AWS runs ~350 data centers globally; US-East-1 in Northern Virginia is the oldest and most critical region despite multi-zone redundancy.
- Supabase had 10+ days of EU-West-2 downtime separately because AWS refused to allocate them additional capacity.
- X (Twitter) was unaffected, giving Elon Musk and DHH public ammunition against AWS cloud concentration.
2025-10-21 · Watch on YouTube
