AWS Outage Takes Thousands of Websites Offline for Three Hours (cnbc.com)
- Reference: 0179835684
- News link: https://tech.slashdot.org/story/25/10/20/140248/aws-outage-takes-thousands-of-websites-offline-for-three-hours
- Source link: https://www.cnbc.com/2025/10/20/amazon-web-services-outage-takes-down-major-websites.html
Gaming platforms including Fortnite, Clash Royale and Clash of Clans went offline. Signal confirmed the messaging app was down. In Britain, Lloyd Bank, Bank of Scotland, Vodafone, BT, and the HMRC website faced problems. United Airlines reported disrupted access to its app and website overnight. Some internal systems were temporarily affected. Delta experienced a small number of minor flight delays. By 3:35 a.m. Pacific time, AWS said the issue had been fully mitigated. Most service operations were succeeding normally though some requests faced throttling during final resolution. AWS holds roughly one-third of the cloud infrastructure market ahead of Microsoft and Google.
[1] https://www.cnbc.com/2025/10/20/amazon-web-services-outage-takes-down-major-websites.html
Luckily for us (Score:3, Funny)
Slashdot is hosted on two squirrels and a dead badger running Linux.
Re: (Score:2)
> Slashdot is hosted on two squirrels and a dead badger running Linux.
Wow, +1 for the obscure Lucy Snyder reference!!
[1]http://strangehorizons.com/wor... [strangehorizons.com]
[1] http://strangehorizons.com/wordpress/non-fiction/articles/installing-linux-on-a-dead-badger-users-notes/
It's always DNS... (Score:5, Funny)
It's not DNS!
It's not DNS!
I promise it's not DNS!
Dammit, it's DNS...
Re: (Score:2)
Lol, so been there.
Re: (Score:2)
hey the missing DNS is just that the healtcheck for dynamodb failed, with no valid endpoints, you have no dns
Roblox? (Score:1)
At midnight? Isn't Roblox a kids' game? What are kids - thousands, according to the fine summary - doing awake at midnight on a school night?
Re:Roblox? (Score:4, Informative)
In France, not only was it day time, it's a holiday week for much of the country's school kids.
Re: (Score:2)
Oh duh. You're right. Nevermind.
Re: (Score:2)
Are you somehow under the impression that all of the children in the world live in a single timezone?
Re: (Score:2, Insightful)
Roblox is a game for adults to go and try to meet kids.
It should be investigated and closed. No parents with any sense should let their kids play it.
Re: Roblox? (Score:3)
This is the Internet, not America.
Pathology in numbers (Score:2)
There's a funny effect that happens with mass centralization of services like this - there's a "safety in numbers" instinct that causes mass failures.
When you self-host, you are obviously responsible for managing outages. Part of what you pay for with AWS is outsourcing that blame.
AWS outages are treated like the weather; nobody's fault, It Just Happens. If your DC is down, you must be incompetent.
Some things still broken... (Score:1)
I'm trying to work and it seems like Atlassian is still having problems recovering. "Put it in the cloud!", they said. "It'll be more reliable!", they said.
Re: Some things still broken... (Score:3)
It's true, we never had outages before the cloud. Yet another CSP innovation.
Re: (Score:2)
Searches on amazon.com aren't working for me right now.
L O N G E R !! (Score:2)
I noticed DNS slowness all Sunday. Not enough to break elinks but uncharacteristic. My primary nameserver is in CO, nowhere near MAE-east.
AWS said the issue had been fully mitigated (Score:2)
Downdetector says AWS is just plain lying.
Cloud hosted JIRA is down as well... (Score:2)
So I don't have to worry about getting any help desk tickets for these issues.
They also say that they are "moving mountains" to resolve my issue with Confluence being down. Funny... I don't remember AWS having issues with a mountain landing on their us-east-1 data centers. Just fess up and tell us what the real problem is? It's not like we don't already know.
AWS Outage (Score:4, Informative)
Oh, I know. Believe me.
Only select services were--and continue to be--affected, however. The root cause seems to be related to a failure in DNS resolution for DynamoDB affecting the entire us-east-1 region. This caused all kinds of AWS internal APIs to fail which relied on this service.
Yet pointy hairs... (Score:2)
... will still believe that Cloud is somehow magical and immune from outages despite plenty of examples to the contrary, particularly from Azure in the last few years.