AWS US-EAST-1 sign-ins and sign-ups fail for over three hours
AWS gave no cause for the Sunday outage, which hit IAM Identity Center, new account sign-ups and network load balancers, and left administrators without a workaround for part of the incident.

Amazon Web Services' US-EAST-1 region returned sign-in, sign-up and network connection errors for more than three hours on Sunday, November 15, 2026, according to a report by Pasquale Pillitteri. The errors began at 3:05 a.m. Pacific time, which is 12:05 CET, and the final network issue was resolved at 6:21 a.m., or 15:21 CET.
The failure unfolded in two phases. Authentication was restored at 5:20 a.m. (14:20 CET), but a narrower network issue involving a subset of Network Load Balancers in the use1-az1 Availability Zone persisted until 6:21 a.m.. AWS engineers were engaged automatically at 3:10 a.m., five minutes after the first errors. The timeline comes from AWS notices as reconstructed by Cyberpress, according to the report.
AWS Support said on X that it was investigating connectivity issues in US-EAST-1 and that "IAM Identity Center sign-in requests and AWS SignUps are also impacted at this time". The company gave no cause in its first notice or its closing message.
The report said AWS's list comprised eleven services, among them IAM Identity Center, IoT Device Management, NAT Gateway, and Sign Up, as well as Step Functions, CloudWatch, Connect, Elastic Load Balancing, OpenSearch Service, VPC Lattice, and Kiro. At 5:15 a.m., AWS said affected customers had no workaround for authentication. Those who logged into Kiro via a social login were not affected, while sign-ins through Identity Center failed, and attempts to create new accounts also produced errors. For administrators, the impact was worse than a compute failure because IAM Identity Center acts as the gatekeeper; if the gatekeeper is unresponsive, the systems behind it may be fine but cannot be reached.
For call centers built on Amazon Connect with agents admitted through Identity Center, the practical effect was workstations without access during the European working day. For the remaining network problem, AWS advised customers who could not wait to create a new Network Load Balancer with targets spread across multiple Availability Zones. The company asked customers to retry failed requests but gave no other immediate guidance for those locked out of the Console.
The company used two closing times: it gave 6:10 a.m. as the end of the main window and 6:21 a.m. for the last load balancer issue. Individual customers could have seen sign-in failures for about two and a quarter hours or connection problems for more than three hours. AWS has not said whether the sign-in problem and the load balancer problem shared the same origin, and no notice mentioned an attack or compromised credentials.
US-EAST-1 is AWS's first region, opened in 2006, and some global functions depend on its endpoints. The outage follows a larger October 2025 failure in the same region tied to a DNS problem with DynamoDB that affected more than 75 services for over fifteen hours, though AWS has not linked the two events. The failure landed on a Sunday, in the middle of the day for European companies and overnight for California customers.
Sources: Pasquale Pillitteri
Kai Brennan covers cloud providers, outages and vendor pricing for Five Nines. He writes from status pages, vendor announcements and published reporting.
Corrections. Report an error to corrections@thefivenines.com. We list substantive corrections on the corrections page.