Amazon says Web Services are recovering after outage hits millions of users – as it happened
Amazon Web Services suffered a major outage that disrupted millions of users across gaming, social media, banking and Amazon's own retail site. The company has posted incremental recovery updates, while regulators and customers demand answers.
On 8 October 2023 Amazon Web Services (AWS) experienced a widespread outage that knocked out or degraded dozens of high‑profile applications, from Fortnite and Roblox to SnapChat, Lloyds Bank and Amazon’s own shopping platform. By early afternoon Pacific Time the company reported that recovery was progressing across all AWS services, with Lambda invocation errors fully restored and EC2 instance launches gradually returning to normal.
What went wrong and which services were hit?
The disruption originated in the US‑East‑1 (Northern Virginia) region, where an internal subsystem that monitors network load balancer health failed. The failure cascaded through core services such as DynamoDB, SQS, EC2, and Lambda, causing API errors, throttling of new virtual server requests and intermittent connectivity problems. As a result, popular consumer‑facing platforms reported errors or slow performance. Gaming services Fortnite and Roblox, social apps SnapChat, Signal and Duolingo, and productivity suites like Adobe Creative Cloud and Microsoft 365 all displayed outages or degraded performance. Even Amazon‑owned services such as the Ring doorbell system and the main Amazon.com retail site displayed the generic “Sorry, something went wrong on our end” message featuring the company’s mascot dog.
In the United Kingdom, the outage affected Lloyds Bank, Halifax, Bank of Scotland, HM Revenue & Customs, and the National Rail website, prompting the House of Commons Treasury Committee to write to Economic Secretary Lucy Rigby asking why AWS has not been designated a “critical third party” under the UK’s financial‑services regime. Similar disruptions were reported in Australia and Canada, with Downdetector logging more than 6.5 million user reports worldwide.
Amazon’s response and recovery steps
Throughout the day AWS issued multiple status updates. At 10:45 am PDT the company announced that Lambda invocation errors had fully recovered and that it was scaling up SQS queue polling to pre‑event levels. By 12:15 pm PST the outage status was downgraded from “degraded” to “impacted,” and AWS confirmed that recovery was visible in several Availability Zones within the US‑East‑1 region. Engineers also implemented throttling limits on new EC2 instance launches to reduce load on the network and applied mitigations to the load‑balancer health‑check subsystem.
Despite the technical chaos, Amazon continued to market its upcoming AWS re:Invent conference scheduled for 23 October. An email sent to customers during the outage reminded them to register for the event, which will showcase new AI and machine‑learning services. The juxtaposition of promotion and crisis drew criticism from observers who argued that the company was downplaying the severity of the incident.
Broader implications and regulatory scrutiny
The outage reignited long‑standing concerns about the concentration of internet infrastructure in a handful of cloud providers. Cori Crider, executive director of the Future of Technology Institute, warned that the UK is “dangerously overexposed to foreign Big Tech monopolies.” Legal experts noted that AWS customers are unlikely to receive compensation for downtime, underscoring the need for clearer contractual protections.
In the UK, Treasury Committee chair Dame Meg Hillier asked three key questions: why Amazon has not been classified as a critical third party, when such designation might occur, and whether the US‑centric nature of the outage poses a systemic risk to British financial services. The committee’s letter reflects growing political pressure to treat cloud providers as essential utilities, similar to electricity or water services.
What’s next for affected users and businesses?
Most services began to return to normal by late morning UK time, but a subset of applications—including Spotify’s Merch Hub, Life360, and certain Adobe services—continued to report intermittent errors. Companies that rely on AWS are advised to monitor the AWS Service Health Dashboard for real‑time updates and to consider multi‑cloud or hybrid‑cloud strategies to mitigate future risk.
Amazon’s share price rose modestly (about 0.7 %) on the day of the outage, suggesting that investors view the incident as a temporary hiccup rather than a fundamental flaw. Nonetheless, the event serves as a reminder that global digital commerce, entertainment and financial services increasingly depend on a single provider’s infrastructure.
As AWS continues to roll out mitigation steps, the next major update is expected around 10 am PDT (6 pm UK time). Stakeholders are watching for confirmation that the underlying load‑balancer health‑check subsystem has been fully restored and that API error rates have returned to baseline levels.
Why it matters
The outage highlights the systemic risk of relying on a single cloud provider for critical services worldwide.
Key points
- AWS outage originated in the US‑East‑1 region due to a load‑balancer health‑check failure.
- Millions of users were affected across gaming, social media, banking and Amazon retail sites.
- AWS reported full recovery of Lambda errors and gradual restoration of EC2 launches by early afternoon.
- UK regulators are questioning why AWS is not classified as a critical third party for financial services.
- The incident fuels debate over cloud concentration and the need for multi‑cloud resilience.
Frequently asked questions
What caused the Amazon Web Services outage on 8 October 2023?
An internal subsystem that monitors network load balancer health in the US‑East‑1 region failed, triggering cascading errors across DynamoDB, SQS, EC2, and Lambda.
Which popular apps and services were impacted by the AWS outage?
Fortnite, Roblox, SnapChat, Signal, Duolingo, Adobe Creative Cloud, Microsoft 365, Lloyds Bank, Ring doorbells, and Amazon.com among others experienced errors or downtime.
How is Amazon responding to the outage?
AWS issued multiple status updates, restored Lambda invocation errors, throttled new EC2 launches to reduce load, and applied mitigations to the load‑balancer subsystem while continuing to promote its re:Invent conference.
What regulatory actions are being taken in the UK?
The House of Commons Treasury Committee has written to Economic Secretary Lucy Rigby asking why AWS is not designated a critical third party under the UK’s financial‑services regime.
What can businesses do to avoid similar disruptions?
Companies should monitor AWS health dashboards, consider multi‑cloud or hybrid‑cloud architectures, and review service‑level agreements for clearer compensation clauses.
