Coinbase released a post-mortem report on the large-scale service outage that occurred on May 7. At the time, major services such as trading, deposits/withdrawals, and order processing were suspended for approximately 8 hours. Coinbase explained that some servers went offline due to a cooling system failure at an AWS data center, and subsequently, the core trading system, the matching engine, could not secure the necessary number of servers for normal operation, leading to a halt in trading. It also stated that an outage occurred in the AWS Managed Kafka (MSK) service, which processes transaction data and order information, delaying recovery. This reportedly caused problems with key functions such as quote provision, fee calculation, and ledger processing. Coinbase stated that although the service should have been designed to operate normally even with a single data center failure, they were not adequately prepared, and plan to proceed with matching engine redundancy and data processing infrastructure improvements in the future.
