Docker experienced significant disruptions due to a widespread outage in AWS’s US-East-1 region on October 20, 2025. Developers worldwide rely on Docker as part of their daily workflow, and we regret the disruption this caused. In this post, we want to provide transparency about what happened, what we have learned, and how we are strengthening our systems for the future.
Beginning October 20, 2025 at 06:48 UTC, Docker Hub, Hardened Images, Scout, Build Cloud, Automated Builds, and Testcontainers Cloud experienced an increase in failure rate when AWS’s largest region, US-East-1 experienced an outage with DynamoDB, EC2, Network Load Balancer, and other AWS services. See the AWS summary of the service disruption for further details. This increasing failure rate led to a complete outage of Docker services across the aforementioned products beginning at 2025-10-20 08:01 UTC.
Starting at 2025-10-20 09:40 UTC, AWS reported progress and partial restoration that resulted in Docker services being partially restored to operation. Full restoration of Docker services was completed by 2025-10-21 09:42 UTC. The complete timeline and user impacts can be seen here.
AWS DynamoDB and EC2 APIs begin failing, causing degraded performance across Docker Hub, Build Cloud, Testcontainers Cloud, and other related services.
AWS STS begins failing with cascading failures across AWS services
Degradation of Docker services increases.
Users experience widespread, increased intermittent failures across all requests
All services unavailable
Users continue experiencing high error rates across all Docker services – over 90%
Docker Hub recovery begins – error rate less than 20%
Docker Hardened recovery begins – error rate less than 20%
AWS EC2 recovery begins with throttling in effect
Docker Scout recovery begins
Docker Offload recovery begins
Docker Build Cloud recovery begins
Docker Testcontainers Cloud recovery begins
Automated builds remain unavailable
Docker Hub and Scout fully recover
Docker Build Cloud and Testcontainers Cloud seeing improvements – error rate ~50%
Automated Builds remain unavailable
Docker Build Cloud, Offload, & Testcontainers Cloud fully recover
Automated Builds remain unavailable
Automated builds fully recover
All services operational
Ongoing Monitoring
All Docker systems are currently operational and we continue to monitor the status of our infrastructure. For real-time operational details, visit our status page where you can subscribe to notifications.
We take Docker Hub’s reliability seriously and understand its critical role in development workflows worldwide. Among Docker’s services, Hub’s registry operations, especially image pulls, are the most heavily used and the most essential to keeping developer workflows moving.
Our first priority is ensuring Docker Hub remains available even in the event of a regional cloud failure. To that end, our immediate focus areas include:
Caching strategies: Expanding and optimizing cache layers to reduce the blast radius of upstream failures, ensuring customers can continue accessing frequently used images even during partial outages.
Multi-region resilience: Enabling regional redundancy for Docker Hub’s read operations, beginning with image pulls. This will allow pulls to continue seamlessly even if a single cloud region experiences disruption. We are also exploring approaches to extend these capabilities to write operations such as image pushes, which involve significantly more complexity across regions.
The Docker community depends on Hub’s reliability, and we take that trust seriously. We are committed to learning from this event so that Docker Hub remains a dependable foundation for developers everywhere.
Quelle: https://blog.docker.com/feed/
Published by