Coinbase details 8-hour outage from May AWS failure, vows improvements

Coinbase has released a post-mortem report on a major service outage that occurred on May 7, which suspended key services like trading, deposits, withdrawals, and order processing for approximately eight hours. The company explained that a cooling system failure at an Amazon Web Services (AWS) data center took some servers offline. This prevented its core matching engine from securing enough servers to operate, halting all trades. Recovery was further delayed by a failure in the AWS Managed Kafka (MSK) service, which processes trade data and order information, leading to issues with price quotes, fee calculations, and ledger processing. Coinbase acknowledged that its systems should have been designed to withstand a single data center failure but were not sufficiently prepared. The exchange announced plans to improve its infrastructure by implementing matching engine redundancy and enhancing its data processing systems.