Svmuu News Coinbase released a post-incident review report on the July 14 service outage, stating that the incident was caused by the misconfiguration of a critical network component triggered by a routine configuration update. The outage lasted approximately 50 minutes, during which user funds remained secure at all times.
Coinbase The company stated that at 12:34 p.m. Beijing Time on July 14 (12:34 p.m. Eastern Time), it deployed a routine configuration change to the shared production Kubernetes cluster hosting its core infrastructure services. This update was originally intended to facilitate a migration to a new service deployment model aimed at improving system performance and reliability.
However, because a resource name conflict was not detected by pre-production testing mechanisms, this configuration change unexpectedly affected Kubernetes resources related to the Istio Ingress Gateway within the cluster, rendering this network component unavailable. Approximately 3 minutes later, all inbound traffic to the cluster was interrupted, preventing internal services from accessing multiple infrastructure components.
Coinbase It was noted that multiple business lines were affected because a large number of asynchronous workflows rely on these infrastructure services, including processes such as transaction settlement, fund transfers, and bank card transaction authorizations. Affected services included:
Retail users were unable to complete off-chain transactions, top-ups, and withdrawals; some transactions in progress appeared to be stuck but were completed after service restoration;
Coinbase Card debit card transactions temporarily failed; credit card payments were unaffected, but some card management features were unavailable;
Coinbase DEX’s on-chain exchange services on Base and Solana were suspended;
Coinbase Exchange and Prime institutional clients experienced failed or delayed transfers and settlements;
Coinbase Developer Platform clients were unable to complete account openings, fund transfers, and deposits.
Coinbase According to the report, service was restored at 1:20 p.m. Eastern Time, and the incident was resolved by 1:23 p.m. The system then began processing the backlog, with most services quickly returning to normal and the remaining queue being processed within the following hours. During the recovery process, Coinbase encountered two additional challenges.First, the deployment tools themselves relied on the affected gateway to operate, preventing the standard rollback process from being executed and creating a circular dependency where “the failure affected the remediation tools.” Second, while the emergency recovery process (break-glass access) was effective, permission verification and manual review procedures extended the recovery time.
Coinbase stated that this incident did not involve any security risks to assets, but it highlighted the need for continued optimization of large-scale financial infrastructure in terms of automated deployment, system dependencies, and disaster recovery mechanisms. The company will further strengthen system resilience to reduce the likelihood of similar incidents occurring again.