Coinbase has published a post-mortem report detailing the extensive service outage on May 7, 2026, which lasted approximately 8 hours, with full recovery taking about 12 hours. The outage was caused by the failure of multiple cooling units in the AWS us-east-1 region's Availability Zone use1-az4, leading to thermal shutdowns of server racks. This resulted in EC2 instances and EBS volumes going offline, severely impacting Coinbase's trading, deposits, withdrawals, and core services. The incident highlighted vulnerabilities in Coinbase's infrastructure, particularly its reliance on a single AWS data center for its trading matching engine, which lost quorum during the outage. Additionally, AWS-managed Kafka (MSK) experienced a control plane failure, disrupting various systems. Coinbase plans to enhance its cross-region architecture, improve disaster recovery protocols, and collaborate with AWS to address these issues and prevent future occurrences.