Worldwide AWS Outage
Worldwide AWS Outage
After experiencing trouble for hours on Monday, services such as Snapchat, Canva, Fortnite and more are now operating normally. This problem was caused by an outage of Amazon’s computing services unit, Amazon Web Services.
On Monday, October 20 2025, AWS the cloud-computing arm of Amazon suffered a widespread global outage affecting thousands of applications, websites and services that rely on its infrastructure.
The troubles began in the early hours of the day (US East-Coast time) in the US-EAST-1 region (Northern Virginia) and quickly cascaded outward. Affected services ranged widely including consumer apps, gaming platforms, voice assistants, banks, and enterprise tools.
Cause & timeline
AWS determined that the disruption stemmed from DNS resolution issues (Domain Name System) in the US-EAST-1 region specifically involving one of its key database services (DynamoDB) and the internal monitoring subsystem for network load balancers.
To quote AWS: “Between 11:49 PM PDT on October 19 and 2:24 AM PDT on October 20 … increased error-rates for multiple services in the US-EAST-1 Region … the event was the result of DNS resolution issues for the regional DynamoDB service endpoints.”
Experts note that this appears to have been an internal error rather than a malicious cyberattack.
Timeline
- 07:11 UTC (≈ 3:11 AM ET) — AWS issues first alert of “increased error-rates” in US-EAST-1.
- Hours into the event — outage reports spike across multiple services, regions. Universities, banks, gaming platforms report failures.
- 6:00 PM ET — AWS says services “returned to normal operations”.
What was affected
- Consumer apps & social: Snapchat, Fortnite, Duolingo, Signal, WhatsApp all reported disruption.
- Gaming & entertainment platforms: Epic Games/Epic Games Store, Roblox, etc.
- Smart-home / voice devices: Alexa, Ring cameras.
- Financial services & banks: In the UK, Lloyds Banking Group (and its subsidiaries) and HM Revenue & Customs reported issues.
- Education & corporate tools: University systems, platforms like Canvas, Smartsheet, etc.
Why it matters
- Concentration risk: AWS holds a very large share of the global cloud-infrastructure market (estimates ~30 % for AWS alone) which means when AWS falters, the ripple is massive.
- Global interconnectedness: A technical fault in one region in the U.S. can affect users and businesses around the world—including you in India.
- Business & operational impact: Even though the outage lasted only hours in many cases, delayed transactions, failed payments, disrupted workflows and lost productivity can cascade.
- Resilience & architecture lessons: Many experts argue this is a wake-up call for companies to diversify cloud providers, build multi-region redundancy, monitor dependencies and avoid single-points-of-failure.
The broader picture & aftermath
- The event adds to a history of major AWS outages (2020, 2021, 2023 etc.) mostly tied to US-EAST-1.
- Regulators / governments are taking notice: In the UK, the outage sparked questions about whether AWS should be designated a “critical third-party” service provider subject to tighter oversight.
- Financial and reputational cost: The full cost is still being tallied, but for businesses the indirect costs (lost productivity, customer trust, manual workarounds) may exceed the immediate downtime.
- Technical takeaway: Even the cloud giants are vulnerable. The infrastructure that powers much of the internet is only as strong as its least-resilient link.
The 20 October 2025 AWS outage served as a stark reminder that even the backbone of the modern internet — massive cloud infrastructure serviced by a handful of providers can fail. For hours people found apps not working, websites unreachable, banks glitching, smart devices offline.
While AWS says the issue is now resolved, many of the downstream effects (in business processes, data-backlog, customer complaints) may linger. The event also shines a light on the need for businesses and users alike to consider resilience not just “it works now” but “how do we keep working when it doesn’t?”
