The disruption came during a turbulent period for Coinbase following the company’s quarterly results and announcement that it plans to cut about 14% of its workforce as it strives to become “lean, fast, and AI-native.”
This is “never acceptable”
For hours, traders in crypto markets speculated about the cause of the outage, with users reporting problems accessing accounts, executing trades and moving funds.
Armstrong addressed the incident in an article on X on Friday.
“We experienced an outage at Coinbase last night, which is never acceptable,” he wrote.
According to Armstrong, the disruption began after several coolers at an AWS facility failed, causing a room to overheat and impacting Coinbase systems.
While many Coinbase services were designed to handle outages within a single AWS Availability Zone, Armstrong said the company’s centralized exchange architecture was more exposed because of the way trading platforms are optimized.
“Exchanges have unique architectures that optimize latency and client co-location,” he wrote.
Crypto Still Runs on Big Tech
Armstrong said Coinbase will now revisit some of these infrastructure trade-offs post-release.
“At a minimum, the duration of an outage should be able to be significantly reduced when an AZ move is necessary,” he said.
Coinbase said customer funds remained safe throughout the disruption.
Shortly after Armstrong’s post, the head of the Coinbase platform Rob Witoff shared additional details about the outage, saying outages inside the affected AWS zone disrupted parts of the exchange’s matching engine and Kafka messaging infrastructure.
Witoff said Coinbase’s main exchange systems operate in a single zone to minimize latency, while backup systems designed to isolate outages did not perform as expected during the incident, extending the output and forcing engineers to manually perform disaster recovery procedures.
Image via Shutterstock
