Between October 19 and 20, an outage at Amazon Web Services (AWS) caused service interruptions affecting approximately 2,000 AWS client companies, according to the website[1], compiles reports of service outages submitted by users. The outage was felt, directly or indirectly, by millions of users of services such as Snapchat, Signal, Roblox, and banking apps.
AWS explained in detail the cause of the outage[2]. The outage was caused by a faulty DNS record linked to DynamoDB, a serverless database provided by AWS and used to store its customers’ data. Such a database has a complex DNS architecture, which is managed by automated systems. Amazon states that a service such as DynamoDB “maintain hundreds of thousands of DNS records to operate a very large heterogeneous fleet of load balancers in each Region. Automation is crucial to ensuring that these DNS records are updated frequently to add additional capacity as it becomes available, to correctly handle hardware failures, and to efficiently distribute traffic to optimize customers’ experience“[3].
The malfunction encountered at DynamoDB also impacted the Elastic Compute Cloud (EC2) service, which allows AWS customers to set up cloud-hosted applications and easily increase or reduce the resources needed to operate them. This solution allows AWS customers to avoid investing in in-house equipment and infrastructure, which would be much more expensive.
As a consequence, a faulty DNS record on an AWS service had a domino effect, causing service interruptions that affected thousands of applications and services. One example is EightSleep connected beds: the AWS outage prevented customers of this high-end connected bed brand from using their dedicated app to adjust the angle or temperature of their mattresses. In a post on social media platform X, the EightSleep CEO swiftly announced an update that would allow the beds to be used offline via Bluetooth[4].
While the example of EightSleep may seem amusing, it should nevertheless give us pause for thought. AWS is a major player in the cloud market, with around 30% market share. Players such as AWS, Microsoft Azure (whose services have also been disrupted in recent days[5]) and Google Cloud together accounted for 63% of the market in the second quarter of 2025[6].
The DNS system was built on the principle of decentralization, but we are seeing a concentration of hosting for thousands of services and platforms used in our daily lives among the Providers mentioned above. This concentration is driven by a need for performance and cost reduction. Hosting a platform, service, or online application on your own is expensive in terms of equipment, infrastructure, and maintenance. Providers such as AWS enable their customers to benefit from a powerful and resilient cloud infrastructure at a much lower cost than an in-house solution, with equivalent performance.
However, as we saw between October 19 and 20, this concentration creates de facto vulnerabilities: a malfunction at a provider such as AWS leads to the partial or total interruption of thousands of services around the world, affecting millions of users, sometimes in very tangible ways, as in the case of EightSleep beds. This outage, with its complex origins, reminds us of the importance of implementing DNS redundancy, particularly when a DNS malfunction can cause a “domino effect” such as the one we saw between October 19 and 20
AWS quickly determined the cause of the outage, restored the affected service, and announced that it was working to improve its tools to prevent the situation from recurring.
The IP Twins teams support you in managing your domain name portfolio and DNS zones through your dedicated account manager.
Notes
[1] “User reports indicate problems at Amazon Web Services” : downdetector.com.
[2] “Summary of the Amazon DynamoDB Service Disruption in the Northern Virginia (US-EAST-1) Region” : aws.amazon.com.
[3] Ibid.
[4] Xcancel.com : https://xcancel.com/m_franceschetti/status/1980419272766583262.
[5] Microsoft says it’s recovering after Azure outage took down 365, Xbox, and Starbucks : theverge.com.