AWS Down Today? What’s Really Behind the December 5, 2025 Internet Outage

On December 5, 2025, huge parts of the internet briefly broke for millions of users worldwide. Trading apps stalled, AI tools wouldn’t load, gaming platforms crashed, and social media feeds filled with screenshots of 500 errors and the familiar question: “Is AWS down today?”

The short answer: No — Amazon Web Services (AWS) is not experiencing a new global outage today. The disruption is being driven by Cloudflare, a separate internet infrastructure provider that suffered its second major outage in less than a month. ^[1]

Below is a clear breakdown of what’s happening, why everyone keeps blaming AWS, and how to check the real status of your cloud services.

Is AWS Down Today (December 5, 2025)?

As of publication on December 5, 2025, there is no new large‑scale AWS outage similar to the one that hit in October:

The AWS Health Dashboard, which lists service disruptions by region, is currently showing normal global status updates with no fresh widespread incident notices. ^[2]
A live outage blog tracking today’s internet problems explicitly notes that “AWS is not down” and points to AWS’s status page showing all servers running normally. ^[3]
AWS is in the middle of its annual AWS re:Invent 2025 conference in Las Vegas (Dec 1–5), with no conference-related announcements of an active cloud failure. ^[4]

If your app runs on AWS but is still broken right now, it’s most likely because Cloudflare sits in front of it (for DNS, CDN, or security), and Cloudflare’s outage is blocking users from reaching an otherwise healthy AWS backend.

What Actually Went Down: Today’s Cloudflare Outage

Today’s chaos stems from Cloudflare, one of the world’s largest providers of content delivery, DNS, and security services.

A brief but global shock

A major incident on Cloudflare’s network this morning caused a widespread internet outage, knocking out access to a large number of websites and apps around the world. ^[5]
Cloudflare confirmed issues on its status page around 08:56 GMT, and by roughly 09:12 UTC a fix had been implemented, with traffic gradually returning to normal as engineers moved the incident into “monitoring” mode. ^[6]
One live report estimated that roughly 20% of all websites were affected at the peak of the disruption, creating a visible “ripple” across the internet. ^[7]

Who was affected?

Reports from multiple outlets and monitoring services list a long roster of impacted platforms:

Video & collaboration: Zoom, Google Meet. ^[8]
Gaming & entertainment: Fortnite, Roblox, League of Legends, Spotify, Disney+ and other major services saw outages or severe errors. ^[9]
Social & professional networks: X (Twitter), LinkedIn, Substack and other content platforms showed 500 errors or became unreachable. ^[10]
Financial & trading apps: Zerodha, Angel One, Groww and other brokers reported login and trading issues, leaving users unable to place orders or see live data. ^[11]
AI services: Claude, Perplexity and other AI-powered tools stopped responding for many users. ^[12]

Users mostly saw HTTP 500 “Internal Server Error” pages, a typical sign that something is wrong deep in the service’s infrastructure rather than on their own devices. ^[13]

Why People Keep Asking “Is AWS Down?” Every Time the Internet Breaks

The confusion today isn’t random. The internet is still processing the shock of a massive AWS outage on October 20, 2025, which genuinely did take big chunks of the web offline.

The October 2025 AWS meltdown

On October 20, 2025, AWS’s US‑EAST‑1 region in Northern Virginia suffered a major disruption that lasted over 15 hours and affected hundreds or thousands of companies worldwide. ^[14]

Key points from post-incident analyses and reporting:

The incident began as a DNS race condition affecting DynamoDB endpoints, cascading into failures across dependent services. ^[15]
AWS later explained that an internal subsystem that monitors the health of its network load balancers misbehaved, breaking the Domain Name System paths used by critical APIs. ^[16]
As a result, a wide range of platforms including Snapchat, Coinbase, Robinhood, Venmo, Amazon.com, Prime Video, Ring, Alexa, education platforms like Canvas, and many others experienced outages. ^[17]
Several analyses note that this was at least the third major US‑EAST‑1 meltdown in five years, raising questions about architectural risk in that region, which is often the default for many AWS customers. ^[18]

Because that October disruption was so visible and so recent, people have effectively built a reflex: when big apps stop working, they assume “AWS must be down again.” Commentary pieces have already framed recent Cloudflare and AWS failures as repeated blows to the perception of cloud reliability. ^[19]

Today, that reflex is being triggered again — even though the underlying problem is different.

How Cloudflare and AWS Interact in Real‑World Architectures

A key reason for today’s confusion is that Cloudflare and AWS often work together in front of the same applications:

AWS typically hosts the core infrastructure — compute, databases, storage, queues and internal APIs.
Cloudflare often sits “in front” of that stack, handling DNS, caching (CDN), TLS termination, DDoS protection and WAF (web application firewall).

When Cloudflare is down:

DNS lookups may fail, so your browser can’t even find the AWS-hosted endpoint.
Even if AWS services are perfectly healthy, users may see 500 errors or complete timeouts, because requests hit Cloudflare first and die there. ^[20]
From the outside, it feels like the app — and therefore AWS — is broken, when in reality the bottleneck is the edge provider.

This layered structure means outages at either layer — Cloudflare at the edge or AWS at the core — can make the internet feel like it’s collapsing, even if the other provider is fully operational.

What We Know Right Now (December 5, 2025)

Putting it all together, here’s the clearest snapshot of the situation:

Cloudflare suffered a major global outage this morning, briefly disrupting access to a significant share of websites worldwide. ^[21]
The issue has been mitigated, with Cloudflare reporting a fix and moving to monitoring; third‑party trackers show error reports dropping sharply. ^[22]
AWS is not currently experiencing a new global outage. Media coverage of today’s incident specifically notes that AWS’s status page lists all services as running normally, and AWS’s own health dashboard has no fresh, mass‑scale disruption listed for December 5. ^[23]
Ongoing performance issues you see in apps hosted on AWS are most likely side‑effects of Cloudflare’s outage, residual routing changes, or localised problems rather than a fresh AWS meltdown.

How to Check If AWS Is Really Down

With so many moving parts in modern internet infrastructure, it helps to have a checklist you can follow before panicking.

1. Check the AWS Health Dashboard

Go to the official AWS Health Dashboard and look at:

Global status: Is there a prominent banner about a regional or multi‑service incident?
Your region: Check US‑EAST‑1, EU‑WEST‑1 or whichever region your workloads use.

Large incidents, like the October 20 US‑EAST‑1 outage, are formally documented with start times, affected services, and progress updates, and remain visible even after resolution. ^[24]

2. Cross‑check with independent reporting

If something big is happening, it usually appears fast in:

Reputable tech and business media
Live outage blogs (for example, those that tracked today’s Cloudflare disruption and explicitly ruled out a current AWS outage) ^[25]

If coverage is focused on Cloudflare — and not a fresh AWS status post — that’s a strong signal about who is really at fault.

3. Look at which services are failing

Patterns matter:

If you can’t reach a wide variety of sites across different industries and clouds, and many show 500 errors, it’s often an edge/CDN/DNS issue (Cloudflare, another CDN, a major DNS provider). ^[26]
If failures cluster around apps heavily known to run on AWS US‑EAST‑1 — especially if official AWS messages mention that region — then you may be seeing a genuine AWS outage. ^[27]

4. For engineering teams: diversify

After the October AWS outage and repeated Cloudflare incidents, cloud experts and analysts are doubling down on resilience advice:

Don’t treat a single cloud region or a single edge provider as your only lifeline.
Consider multi‑region or multi‑cloud architectures, and at least multi‑CDN or backup DNS providers, especially for mission‑critical services. ^[28]

Lessons from the October 2025 AWS Outage

Today’s Cloudflare incident is happening in the shadow of what some have dubbed “The Great AWS Outage of 2025.” ^[29]

In October, the AWS event showed:

Single-region concentration risk: Dependency on US‑EAST‑1 as the default for many workloads amplified the blast radius when things went wrong. ^[30]
DNS and control-plane fragility: A failure in DNS resolution and health monitoring systems — not physical hardware — was enough to knock out core services worldwide. ^[31]
Multi-industry exposure: Everything from streaming video and gaming to fintech, education and even some government services were affected, with billions in estimated losses from downtime and lost productivity. ^[32]

Today’s Cloudflare outage reinforces the same central message: the modern internet is more centralized — and therefore more fragile — than most people realise.

Wider Context: A Month of Cloud and Edge Turbulence

The December 5 Cloudflare incident is part of a broader pattern over recent weeks:

Cloudflare already suffered another global outage in mid‑November, tied to maintenance and a latent software flaw. ^[33]
After the October AWS meltdown, coverage from analysts and business press highlighted how much of the world’s digital infrastructure depends on a small handful of cloud providers, particularly AWS, Microsoft Azure and Google Cloud. ^[34]
In response to growing concern, Amazon and Google have jointly announced a new multicloud networking service to improve connectivity and resilience between AWS and Google Cloud, allowing customers to more easily connect applications across providers. ^[35]

There’s also a growing ecosystem focused specifically on resilience on AWS itself. Today, for example, PagerDuty was recognized with one of the first AWS Resilience Services Competency designations, highlighting tools and practices designed to keep critical workloads running during incidents. ^[36]

Other AWS‑Related Headlines on December 5, 2025

Even as users Google “AWS down today,” AWS is making news for very different reasons:

BlackRock & AWS partnership – BlackRock announced that its Aladdin investment management platform will be available on AWS, emphasizing secure, scalable, and resilient infrastructure and enabling multi‑cloud flexibility for financial clients. ^[37]
Resilience and incident response – PagerDuty’s new AWS resilience milestone (mentioned above) speaks directly to the rising cost of downtime and the need for better incident management on AWS workloads. ^[38]
AI on AWS in Southeast Asia – Thai real‑estate giant Sansiri has rolled out a generative‑AI concierge and large‑scale invoice automation on AWS infrastructure, while Amazon’s stock ticked down modestly despite the positive tech news. ^[39]

In other words, while today’s outage isn’t AWS’s fault, AWS is still squarely in the spotlight — both as a source of previous disruption and as a platform companies rely on for resilience and advanced AI services.

Quick FAQ: AWS vs Cloudflare on December 5, 2025

Is AWS down today?

No, not at the time of writing. Official status pages and live outage blogs indicate that AWS services are operating normally, and today’s widespread errors are being attributed to Cloudflare’s network, not a new AWS failure. ^[40]

Why do my AWS‑hosted apps still show errors?

If your application uses Cloudflare for DNS, CDN or security, users hit Cloudflare first. When Cloudflare is having issues, requests may never reach your AWS backend, even if EC2, DynamoDB, RDS and other services are healthy.

Did AWS have a big outage recently?

Yes. On October 20, 2025, AWS’s US‑EAST‑1 region suffered a major incident tied to DNS resolution and health monitoring, disrupting services like Snapchat, Coinbase, Robinhood, Amazon.com, Prime Video, Ring, Canvas and many more for much of the day. ^[41]

Why does it feel like “half the internet” goes down every few weeks?

Because huge chunks of the web now rely on a small number of core providers — AWS, Azure, Google Cloud, Cloudflare and a handful of others. When any one of them suffers a bug, misconfiguration or maintenance mishap, millions of users around the world feel it almost instantly. Recent outages at AWS, Cloudflare and others have made that structural fragility painfully visible. ^[42]