TL;DR: On October 20, 2025 Coinbase along with many companies experienced severe service disruptions due to a widespread AWS us-east-1 regional outage. There were varying levels of Coinbase customer impact ranging from being unable to log in to being unable to trade. The total amount of time Coinbase users experienced degraded performance during the outage was 3 hours 17 minutes.
On October 20, 2025 at 2:51AM ET AWS services dependent on DynamoDB experienced widespread failures. This impacted Coinbase’s ability to scale to meet traffic demand, maintain visibility into operational health, and led to periods of services being hard down.
Most Coinbase product lines were impacted:
Login services were unavailable or degraded, preventing users from accessing accounts.
Coinbase trading services experienced periods of unavailability.
Transfers, withdrawals, and deposits were delayed or failed.
Staking, onboarding, market data, and crypto send/receive were impacted.
Core vendors Coinbase relies on were also heavily impacted, slowing our ability to communicate and diagnose impact.
Wave 1 Approximately 3AM to 7AM ET:
AWS DynamoDB disruption: DNS resolution failures for DynamoDB API endpoints caused widespread service errors. Coinbase was unable to request help from AWS as the Support Console was offline.
EC2 instance provisioning issues: After mitigation of the DynamoDB DNS issue, requests to launch new EC2 instances continued to experience increased error rates.
Wave 2 Approximately 11AM to 6PM ET:
Network connectivity issues: Multiple AWS services (ELB, NLB, STS, EKS) and EC2 experienced connectivity problems.
EC2 API Throttling: Prevented ENI provisioning which intermittently blocked Kubernetes worker node creation and pod scheduling.
We took immediate steps to mitigate the impact of the incident as much as technically possible. Coinbase engineers were paged at 3:08AM ET, and all service owners participated to reroute traffic so they could be used by customers. This work included targeted mitigations according to our standard recovery procedures such as: disabling auto cluster consolidations, locking deploys, load shedding non-critical services, and manually assigning our fixed capacity pool to remaining services.
All Coinbase services were restored at 6:45pm ET after AWS implemented fixes to multiple systems that Coinbase relies on, in particular, EC2’s network state propagation.
The work done by engineers in the first wave prevented significant impact during the second wave of this widespread outage. To be better prepared in the future, we are exploring all options, including reviewing our regional deployment strategy to implement immediate, and long term fixes to reduce the impact of these types of outages.
Coinbase has entered an agreement to acquire The Clearing Company, a prediction markets company innovating at the frontier of regulated, onchain markets. The team brings deep expertise that will help power and scale prediction markets on Coinbase.
