AWS us-east-1 S3 incident caused by mistyped command

The author argues that while per-site data centre outages are becoming rarer, individual failures are growing much larger and more disruptive because critical economic dependencies are increasingly concentrated in fewer systems. Drawing on data from the Uptime Institute and cloud incident trackers, the analysis highlights that configuration changes, DNS or identity single points of failure, and physical risks continue to recur. The author forecasts a high probability of another major global outage within the next twelve months. Open article →