Michał "rysiek" Woźniak · 🇺🇦
and 1 other
boosted
You can't be DDoSed if your site is down!
I got quoted in The Guardian again, I guess CloudFlare must have been down or something? :blobcatcoffee:
> “These companies have become too big to not fail. And because they handle so much traffic, when they do fail, this immediately becomes a massive problem”
Jenniferplusplus
and 1 other
boosted
There is a book called Normal Accidents:
https://en.wikipedia.org/wiki/Normal_Accidents
The author puts forth three rules that define systems susceptible to catastrophic accidents which are completely to be expected:
- the system is complex
- the system is tightly coupled
- the system has catastrophic potential
This describes huge cloud providers like CloudFlare very well, and specifically describes the last outage very well.
CloudFlare, AWS, Azure, GCP are simply Too Big Not To Fail.
Magnus Manske
and 6 others
boosted
AWS, Microsoft Azure, and CloudFlare – services run by gigantic corporations with endless supply of money and talent – all experience catastrophic, global failures that take innumerable other services down with them within 30 days. :blob0w0:
Meanwhile Wikipedia just keeps chugging along, globally stable and reliable as always. :blobcatlove:
And yes, Wikimedia Foundation runs a pretty complex infrastructure:
https://wikitech.wikimedia.org/wiki/Wikimedia_infrastructure
nullagent
boosted
> In this specific instance, the Bot Management system has a limit on the number of machine learning features that can be used at runtime.
> (…)
> When the bad file with more than 200 features was propagated to our servers, this limit was hit — resulting in the system panicking.
https://blog.cloudflare.com/18-november-2025-outage/
So it's "AI" when it's good for business, "machine learning" when it's bad for business, gotcha! 🤡
alcinnz
boosted
This is, of course, hilarious. If a single traffic spike can bring CloudFlare down for hours, and with it innumerable websites and services, what exactly is the point of CloudFlare?
Yes, I understand that a way smaller traffic spike would probably bring any individual website down if they were all hosted separately. Sure.
But bringing them *all* down at the same time would be impossible. That's resilience through independence.
#CloudFlare is a single point of failure, and it keeps failing.
> In this specific instance, the Bot Management system has a limit on the number of machine learning features that can be used at runtime.
> (…)
> When the bad file with more than 200 features was propagated to our servers, this limit was hit — resulting in the system panicking.
https://blog.cloudflare.com/18-november-2025-outage/
So it's "AI" when it's good for business, "machine learning" when it's bad for business, gotcha! 🤡
AWS, Microsoft Azure, and CloudFlare – services run by gigantic corporations with endless supply of money and talent – all experience catastrophic, global failures that take innumerable other services down with them within 30 days. :blob0w0:
Meanwhile Wikipedia just keeps chugging along, globally stable and reliable as always. :blobcatlove:
And yes, Wikimedia Foundation runs a pretty complex infrastructure:
https://wikitech.wikimedia.org/wiki/Wikimedia_infrastructure
✅ invalid file
✅ deployed globally
✅ crashing the software
CloudFlare just pulled a CrowdStrike?
https://blog.cloudflare.com/18-november-2025-outage/
There is a book called Normal Accidents:
https://en.wikipedia.org/wiki/Normal_Accidents
The author puts forth three rules that define systems susceptible to catastrophic accidents which are completely to be expected:
- the system is complex
- the system is tightly coupled
- the system has catastrophic potential
This describes huge cloud providers like CloudFlare very well, and specifically describes the last outage very well.
CloudFlare, AWS, Azure, GCP are simply Too Big Not To Fail.
✅ invalid file
✅ deployed globally
✅ crashing the software
CloudFlare just pulled a CrowdStrike?
https://blog.cloudflare.com/18-november-2025-outage/
This is, of course, hilarious. If a single traffic spike can bring CloudFlare down for hours, and with it innumerable websites and services, what exactly is the point of CloudFlare?
Yes, I understand that a way smaller traffic spike would probably bring any individual website down if they were all hosted separately. Sure.
But bringing them *all* down at the same time would be impossible. That's resilience through independence.
#CloudFlare is a single point of failure, and it keeps failing.
Yet again I will point out that because "enterprise web development" is cargo-culting these days, a single website might rely on multiple gigantic "cloud" providers at the same time.
We know these exist because when Azure was down, some were blaming CloudFlare. 🤡
There are only so many of these gigantic providers. So if a website relies on CloudFlare, AWS, and Azure simultaneously, over the last month it would have been dead in the water on three different occasions, for hours on end. 👀
You can't be DDoSed if your site is down!