AWS data-center overheating disrupted Coinbase, FanDuel

Overheating at an AWS data center in Northern Virginia knocked out one US-EAST-1 availability zone, halting Coinbase trading for over five hours and blocking FanDuel access.
Overheating at an Amazon Web Services data center in Northern Virginia caused an outage in a single Availability Zone in the US-EAST-1 region, affecting EC2 compute instances and EBS storage hosted in use1-az4.
AWS attributed the disruption to rising temperatures inside one data center and reported that power loss during the thermal event impaired instances and volumes. The company shifted traffic away from the impacted zone and said technicians were working to bring additional cooling capacity online. AWS warned some instances and volumes would remain impaired until full recovery and provided no estimated time for restoration.
Coinbase reported core exchange functions were disrupted for more than five hours and warned some users could experience delayed sends and receives on the Solana network and on ALEO. The exchange later posted: “All markets have been re-enabled for trading on coinbase.com and in the Coinbase iOS and Android apps. Coinbase customers can log in to trade.”
FanDuel acknowledged that users could not access its platform and posted that its team was investigating the technical difficulty.
AWS listed services that had been restored, including AWS IoT Core, NAT Gateway, Amazon Elastic Kubernetes Service, Elastic Load Balancing and Amazon Redshift. The company said other services remained affected at the time of the update, naming Amazon ElastiCache, Amazon Managed Streaming for Apache Kafka, Amazon OpenSearch Service and Amazon SageMaker.
AWS added that mitigation efforts required bringing additional cooling capacity online and recovering hardware in a controlled manner, which extended the recovery timeline. The company did not identify whether a hardware failure, a cooling-system fault or an external factor caused the thermal event.
A widespread AWS outage last year was traced to a DNS issue that disrupted many websites and apps. Many cloud customers distribute workloads across multiple availability zones and regions to reduce the impact of a zone-level failure.







