- An AWS outage took down a wide range of applications Monday morning
- The cloud provider was experiencing DNS issues as well as issues with APIs and spinning up new EC2 instances
- The outage had yet to be fully resolved as of 10:29 am ET
Most of the U.S. was still sleeping when Amazon Web Services’ US-East-1 Region in Northern Virginia went down at around 3 am ET Monday morning and took much of the internet with it. But despite efforts to fix the problems behind the outage, AWS said multiple services were still experiencing issues into the early afternoon.
AWS seemed to be putting out fires on multiple fronts.
The cloud giant initially reported on its health status dashboard that the outage appeared to be related to “DNS resolution of the DynamoDB API endpoint” in its northern Virginia cloud region. Three hours later it stated that the underlying DNS issue was fully resolved.
Just a half hour later, it noted customers were experiencing issues with new EC2 launches. Despite multiple mitigations, the company said at 9:42 am ET that it continued to see “elevated errors” for the service. EC2 is AWS’ scalable virtual server offering.
Around 10:14 am ET, a new issue cropped up, with AWS noting “significant API errors and connectivity issues across multiple services.” About 15 minutes later it said connectivity was beginning to recover, but it was still working out the root cause of the problem.
As of 1:38 pm ET, Amazon noted some problems persisted but "mitigations to resolve launch failures for new EC2 instances are progressing."
By 6 pm ET, Amazon said the issues had finally been resolved.
While it's hard to pinpoint what exactly went wrong this time, Emil Sayegh, CEO of Cybersheath, told Fierce that initial indicators point to a "control plane failure" in the Northern Virginia region that cascaded through to APIs and DNS services. Cybersheath is a government-oriented cybersecurity and compliance company located in Northern Virginia.
"When a foundational layer falters in that region, the impact zone is massive because so many workloads, including national security systems, are anchored there. It is a reminder that overconcentration remains the cloud’s biggest structural weakness," he explained.
The impact of outages
James Kretchmar, CTO for Akamai Technologies' Cloud Technology Group, told Fierce that incidents like the one Amazon faced Monday highlight "just how much reliability matters, and that the industry still has lessons to learn."
"Often companies build in multiple layers of resiliency to prevent a single failure from cascading into a broader outage and so when disruptions do occur, it means several things went wrong at once, and the safeguards themselves failed, too," he said.
Amazon’s last major outage was in June 2023 and impacted the same region that took down a wide range of applications on Monday. Its woes follow outages on Google Cloud Platform and IBM’s cloud service earlier this year.
Cybersheath's Sayegh said two years without an event is "impressive" but doesn't mean a company like Amazon is immune to incidents. That said, he added it likely won't impact AWS' business all that much.
"No enterprise is abandoning AWS over one outage, but many will revisit architecture plans and uptime clauses," he said. "The real differentiator will be how clearly and transparently AWS explains what went wrong."
This story was last updated 10/21/25 at 9 am ET.