Svmuureports that Coinbase CEO Brian Armstrong posted on platform X, stating that Coinbase experienced a downtime last night. The root cause was a malfunction of multiple coolers in an AWS data center, leading to the computer room overheating. Most systems are designed to be redundant and operate normally against a single AWS Availability Zone failure, but the centralized exchange failed to achieve this redundancy due to its optimization for latency and customer hosting. While it is possible to make the exchange resistant to Availability Zone failures, this would introduce latency and disrupt customer hosting.
In light of this incident, these trade-offs will be reevaluated. At the very least, when it is necessary to migrate Availability Zones, downtime should be significantly reduced. Armstrong thanked the AWS and Coinbase teams for mitigating the issue overnight and stated that a detailed technical summary will be shared later.