#(outages)

…the fandom shield holds strong

AWS services are now pretty much unavailable so we lost a good chunk of the internet. Little FYI btw our entire and i mean ENTIRE internet is being held together by 3 large corporations.

Amazon being one.

Google the second.

And microsoft the third.

First nickle is from the Rogers outage that knocked out the debit system in all of Canada. People couldn't call emergency services if their phone provider was Rogers and I think the government was using Rogers for their internet. It was a shitshow. Rogers just got a slap on the wrist and is still one of the three major companies in Canada. Like, this didn't hurt the telecommunication monopoly in Canada at all.

Did your YouTube go down last night (or early this morning, depending on where you are)? How about your Roku? Gmail? Cloudflare? Major outages of services have become a problem in the last few months. AWS and Azure both suffered large scale DNS errors that took out whole swathes of the internet in October 2025. It made headlines, both within and outside of the industry. Both issues were fixed, although Azure’s took roughly 8 hours and disrupted an entire business day. I reported on both of them (here and here, respectively).

This seems to be different, however. No single entity is reporting the same thing, and CloudFlare CTO Dane Knecht insists that reports on Downdetector were not accurate. He wrote on X: ‘I can personally confirm that Cloudflare is running perfectly fine’. Adding, ‘if every service is ‘down' at the same time, maybe the problem isn't the services. There has to be a better solution than Downdetector’.

The thing is, something was going on. Across numerous platforms. Simultaneously. I took this screenshot at 8:34 pm EST (I’ve edited out the rest of my toolbar for privacy).

In the case of YouTube, two thirds of the reports were problems with the app, while a smaller percentage were with the website. The site released a statement clarifying what was going on via X: ‘An issue with our recommendations system prevented videos from appearing across surfaces on YouTube (including the homepage, the YouTube app, YouTube Music and YouTube Kids)’. They claimed the homepage was back up and running and that they were continuing to work on a full fix.

But they weren’t the only ones with a spike in outage reports, as you can see. I admittedly didn’t look at the details of the rest, but in terms of claiming that Downdetector isn’t an authority for checking the status of services...well, okay, it’s not. But it is a tool. How do people think metrics are collected? Yes, checking individual site pages for status is the most reliable way to get information, but Downdetector’s entire purpose is to collate that information in one place at a glance.

Another X user, replying to Dane Knecht’s post, said that they wished people understood that ‘downdetector works by just… tallying the number of people that check downdetector’. That’s rather reductive. Reports are tallied by actively reporting them. You know, clicking on the prompt that asks if you’re having a problem with a service, and what kind of problem that is. Those are the data points that make up those spike graphs. You can see how many there are by hovering over them. We aren’t talking about an insignificant sample size leading to bias, reports were in the thousands or more. According to NDTV Profit, over 2.8 million users reported an outage in the US alone. They weren’t wrong; there was a problem.

Thus far, every article and blog post I’m seeing is focusing on YouTube or CloudFlare, but they were not the only ones with reported issues. So many services suffering an outage simultaneously and for the same length of time is indicative of a larger problem. Or at least of a deeper root cause, since a large proportion were related to one company. And in fact, Google Trust Services did report a security rollout that would prevent issuance from occurring. In plain English, this means site connections were stopped due to a temporary lack of digital identification for devices. A failsafe, in other words, so that private data is not publicly exposed.

So what went wrong? Nothing serious, it appears. It just took longer than expected. It’s likely that the number of ‘outages’ was simply a confluence of lag in a security update combined with a coincidental flaw on the most popular video website in the world during a time of day when internet traffic was high. Taken together, they look like a denial-of-service attack. But that’s sort of backwards. A DoS artificially forces lag so nothing gets through. Lag exists without being malicious. When a device or program hits traffic capacity, it will stop processing until it can prioritize again (see my report on kernels for the nitty gritty). Hopefully that’s all this was, a perfect storm of incidents adding up into a relatively short-lived hassle for users. I’ll be keeping an eye out for any updates, though, in case I’m wrong.

Elon Musk’s X social media platform suffered multiple outages early Monday.

Xbox fandom reaction to an UNPLANNED outage: Mildly outraged, occasional funny haha.

Ao3 fandom reaction to a PLANNED outage: Desperation, Denial, Grief, meme after meme outlining our addiction.