Coinbase says July 14 outage was caused by a configuration error, with no impact on customer funds

Coinbase has published a post-incident review of its July 14 service outage, saying the disruption was triggered by a misconfiguration tied to a routine change deployed to a shared production Kubernetes cluster that supports core infrastructure services. The company said the issue lasted about 50 minutes and did not put customer funds at risk. According to Coinbase, the change was intended to support a migration to a new service deployment model aimed at improving performance and reliability. But a resource naming conflict was not caught in pre-production checks, and the update unexpectedly affected Kubernetes resources tied to the cluster’s Istio Ingress Gateway, knocking that network component offline. Roughly three minutes later, inbound traffic to the cluster stopped, cutting internal services off from several infrastructure components. Coinbase said the outage affected retail trading, deposits and withdrawals, debit card transactions, DEX swaps on Base and Solana, institutional transfer and settlement flows, and parts of its developer platform. The ingress gateway recovered at 13:20 ET, and the incident was mitigated at 13:23 ET. Coinbase also said recovery was slowed by dependency issues in its deployment tooling and by manual checks in its emergency access process.