Yandex outage: data shows that instead of going dark, malicious bots moved
blog.cloudflare.comBuried lede:
>across Cloudflare's network, there was no evidence that the strikes knocked social media manipulation offline or otherwise disrupted malicious activity on the Internet.
This directly refutes the popular memes going around that russian trolls magically went offline because of the attack.
Does Cloudflare have much if anything to do with X/Twitter and Facebook/Instagram and TikTok, or even Reddit?
Presumably these bots make up a very small percentage of traffic in comparison to how much they make up in terms of comments.
And I doubt they have visibility on the most popular social networks.
The reason these bots are able to succeed is because the average person comments and votes so little that there's huge volatility between a power user and an average user. Power users are also not that rare, so it's hard to determine who's a power user and who's a bot (though most power users probably are bots).
You can make up a significant portion of comments and votes (especially on a topic) with a VERY small amount of the total viewing traffic.
Yeah weird thing to purport that feels extremely misleading. Suppose one way to find out is to look at accounts that were suspected of high bot activity and see if they got the same amount of engagement as the weeks before the strikes.
I don't really trust Cloudflare to be a good steward when discussing this since they have a sordid history in protecting hate.
>Yeah weird thing to purport that feels extremely misleading. Suppose one way to find out is to look at accounts that were suspected of high bot activity and see if they got the same amount of engagement as the weeks before the strikes.
But nobody did that. All we have is a bunch of people on twitter/reddit passing around the same pixelated screenshot from cloudflare, using it to claim that russian bots have been stopped in their tracks.
> But nobody did that.
That's quite the militantly ignorant statement.
https://www.thelondoneconomic.com/politics/reform-figures-se...
They also have a virtuous history protecting love.
That's like saying Google protects hate because people who hate you use Google.
When you are colorblind you end up helping everyone even people that hate you.
X is under Cloudflare (and according to some of my probes may even use some parts of CF stack)
you're hopeless. if cloudflare could detect these bots, there wouldn't be bots.
being able to detect a forest fire doesn't mean we no longer have forest fires
The popular memes that were based on people's assumptions or guesses or stories based on... cloudflare own tweets!
https://x.com/cloudflareradar/status/2108122893028671604
https://x.com/CloudflareRadar/status/2110113211059994638
Which cloudflare also doesn't mention, that their vague and misleading tweet was responsible for these rumors that they are now trying to quash too. Tweets starting October 8th and continuing. That they're getting around to being like, yeah guys that data didn't mean that a week later.
This is how "news" and "facts" work in 2026.
>The popular memes that were based on people's assumptions or guesses or stories based on... cloudflare own tweets!
???
>"Bot traffic" refers to [...], not necessarily activity on social media platforms.
I guess they could have been more explicit and straight up said "no, that doesn't mean russian trolls are offline", but their tweet isn't ambiguous.
The clarifying tweet you're quoting was posted 5 days after the first tweet.
The first tweet doesn't even mention anything about bots.
I don’t see anything vague or misleading about either tweet. I suppose it might be misleading if you don’t know what “bots” means in this context and don’t bother to find out?
I mostly saw the tomshardware.com link being the cause:
https://www.tomshardware.com/tech-industry/data-centers/russ...
The word "disinformation" doesn't show up in the post, only in the URL.
This was definitely getting shared around Reddit and a majority the comments were drawing assumptions around the headline saying 'bots'. Honestly I made the same mistake until about the third time I saw the headline in a feed and thought "wait, why would social bots be coming from Yandex?"
Anyone that actually uses Cloudflare and manages their own bot traffic saw those graphs getting passed around and cringed. It has been an interesting thing to watch though as it was disseminated, echoed, and remixed with all sorts of "evidence" of this or that engagement dropping.
Russian state actors and propaganda networks literally using Yandex's ASN is like the CIA calling you from a payphone in the lobby of Langley.
Now you could say that parts of their infrastructure were hosted in there, and I'm sure it was, but the "bot" traffic narrative was pretty stupid.
Thats a bit if a silly analogy isnt it? It would be more correct to say it would be like the NSA/CIA operating out of google datacenters, which they do, extensively.
Which is exactly why the US gov wanted "American" control of TikTok
They wanted control because the Palestinians cause was winning the tiktok war. Controlling cnn, cbs and tiktok will shape those views.
Additionally,
> Following the reported strikes, bot traffic from Yandex.Cloud fell about 40%. … Cloudflare customer blocks on Yandex.Cloud traffic also fell 70% during the event period, suggesting that proportionately less malicious traffic was even hitting the Cloudflare WAF from this ASN
But the memes keep rolling. 57 channels and nothing on.
Yeah, it contradicted my intuition as a developer. If you take down a data center it isn't gone forever. The development team does not live there. Even if it's poorly backed up they're going to rebuild it.
> If you take down a data center it isn't gone forever.
Given the active sanctions against Russia, as well as high worldwide prices for components like memory and storage, rebuilding these datacenters is going to be an exceptionally slow and costly process. And I can't imagine Yandex is going to be eager to rebuild immediately given the potential for a "double-tap".
> If you take down a data center it isn't gone forever.
AWS abandoned Bahrain as a region completely after data center strikes. All Infra went to the scrap
Well, yeah. They relocated their data and their services elsewhere. It didn't go away.
I doubt Yandex would ever be able to leave Russia, even if it wanted to.
I don’t think many people were believing it was solved forever but just that there was a momentary disruption/decrease while they rebuilt it. But according to this report, even that impact was highly exaggerated.
Counter-argument: https://health.aws.amazon.com/health/status
Both were struck over half a year ago and me-south-01 (Bahrain) appears to be gone permanently, while the data loss in me-central-01 (UAE) is limited to one of the availability zones.
Most people making memes never built multicloud or across geo regions
Rhetorical claims that refute isn't really evidence either, just an attempt to null out the prevailing narrative; SEO spam
Elon getting snooty and Starlink issues in Ukraine are just press releases too; how the fuck do we know what anyone actually did? Too many red herrings troubleshooting systems to assume Starlink isn't just glitched sometimes
While I am firmly in the "Ukraine is the good guy" camp generally both sides of a nation state conflict will lie to both deflect blame and project competence and strength. Will give Zelensky benefit of the doubt until he dead or war ends and if he doesn't step aside legally. But he will latch onto headlines to get his message heard even if Ukraine had nothing to do with it.
I am in a couple political discussion groups and it was really curious how the traditional right-wing posters sort of disappeared right around the same time. There are several that people had suspected of being some type of plant that just were suddenly not posting any more and still haven't come back. No idea if it was related but the timing was so close that it raised a few eyebrows.
Is there any serious analysis of this? Presumably it wouldn't be hard to do. Find some right wing telegram channels/x accounts/subreddits, whatever, then chart their activity. All the evidence I've seen are people passing around the cloudflare chart.
I haven’t seen any, but the people that would have access to that data are probably too busy to care. For example, Political Wire noticed a significant decrease in the right-wing trolls, with one longtime nuisance disappearing completely and the general discussion just being way less divisive. The site is basically run by one guy, and he doesn’t have the time or expertise to do some sort of analysis on this. Also, we aren’t talking about some huge number of people here. It only takes a couple of trolls to ruin an online forum, so if the strikes knocked out only a couple dozen trolls, it could make a noticeable difference in the online discourse while being virtually unnoticeable in the overall data.
For what purpose a handful of trolls would require a datacenter? Or any server infrastructure at all, for that matter.
Not really, Cloudflare is talking about a different type of bot, one that scrapes the web.
Whereas social media users are talking about an entirely different type of bot which is injecting manufactured discourse into people's feeds.
No, it describes that Cloudflare was unable to identify that they have or haven't. The lack of evidence is not proof of absence.
you are basing your assumption on a site that have Zero insight about what they are claiming.
cloudflare have zero capabilities to detect mobile farms using sim cards etc. which is what is used for expensive social media manipulation.
What the hell would mobile sim card farms have to do with Yandex? You don't even care about the truth.
oh, but they did. Measurably. Comment sentiments shifted drammatically following the post knockout period, especially on topics where comments typically get flooded by what is commonly called a troll account. Take videos covering pro-choice content or other content critical of US right wing talking points, which draw such bots like flies to shit. Post knockout? Near nada on new videos. It went from a sizable chunk of comments (~13%<amount<27%), to a fraction of a single %. That's a non trivial shift. Still looking at data, but something certainly did happen, and I think the memes are likely grounded in truth.
Just anecdotally I still see as many bots as always. Embrace the bots, have fun with them.
# nginx rule to Awaken AWS GuardDuty and GCP SCC, maybe get them to fix a few VMs, enable on all non-wordpress sites after testing in dev/qa
location ~* /(xmlrpc|wp-login\.php|wp-admin) { log_not_found off; access_log off; return 301 http://169.254.169.254$request_uri; }I really ought to do something like this rather than just allowing our logs to endlessly fill with requests for wp-admin.
Another version if this if the 301 is too much is to just close the connection then have a daemon or cron job shove their IP into a NFTables ipset or blackhole route them. Be sure to have exclusion elements so one can not block their own gateway or corporate SNAT groups.
location ~* /(xmlrpc|wp-login\.php|wp-admin) { set $ban 1; return 444 ''; } # grep banip /etc/nginx.conf log_format banip '$remote_addr'; # grep banip /etc/nginx/http.d/some-server.conf access_log /var/log/nginx/ban.log banip if=$ban;
<send friend request>
There are different kinds of bots. The ones of interest that would have been using Yandex are those that post political comments or misinformation on social media sites. Bots that are just trying generic hacks are more likely to be much more widely distributed, because not every country is trying to hack the US psyche.
The ones of interest that would have been using Yandex are those that post political comments or misinformation on social media sites.
For those bots I put their networks into an nginx geo file along with detecting other characteristics and set a variable that tells my app to allow them to post, but all their posts go into a moderator review queue. Some use residential proxies but I've learned how to automatically detect the never-before-seen residential proxies almost instantly now even when they are properly instrumenting Headless Chrome. Only I see Статуя Свободы мертва.
I thought it was established that those were search engine crawlers...
Incidentally, this morning I asked Gemini Argon 4 on High to do deep research on the effect of the strikes on russian bot farms and after some hours (it got stuck) it generated the report starting with a surprising (to me at the time) paragraph:
"Rigorous digital forensic scrutiny and network telemetry analysis, however, demonstrate that the purported collapse of Russian social media propaganda bots is an analytical mirage born of misinterpreted network traffic data and fabricated viral graphics."
That's funny, this afternoon I asked Claude to think very hard about your comment, and it said you trust LLMs too much.
I asked my dog and it told me it wanted steak
> Rigorous digital forensic s...
Why does my brain refuse to continue reading after this point?
Most Russian bots have a better reputation for fact-checking than Gemini, mate
This seems like a reasonably high probability sequence of tokens, especially if the temperature is set to allow the first few to be generated. What’s the surprise?
Does anyone have the numbers on social media participation? I don't think anyone thought that their ability to execute media influencing activities was completely shutdown, but it did seem that social media was a lot quieter for a time.
Easy hypothesis would be that they were (probably accidentally) reliant on the news scraping bots to find where to focus responses and such. Would be like realizing a lot of US things do down when AWS' us-east goes down. It isn't literally where everything is hosted. A lot of things have accidental dependencies that go through there, though.
This BlueSky user claims to track social media bots/information warfare accounts and has not seen any drop: https://bsky.app/profile/antibot4navalny.bsky.social/post/3m...
There is certainly information warfare happening, but it doesn’t appear that this is the likely host—or perhaps they have good redundancy.
What I’m far more interested in is what the consequences of Yandex cloud being offline are.
I feel like if us-east-1 was down for a week we’d see all kinds of consequences in everyday services. Are Russians enjoying the same?
Yandex cloud is not as big as AWS.
First of all AWS (even in us-*) most likely have non-US services\data. Yandex is mostly used by russian companies.
Plus, from my personal experience cloud is not as popular in Russia as in the west.
Big players like Sber, Ozon, T-Bank etc have their own datacenters and [may] use Yandex (or cloud.ru for example) for redundancy and\or non-critical services.
Small players use vps or something similar.
Big non-tech players (gas&oil for example) often rent redicated servers or space in a rack for their own hardware.
At least in the West, “space in a rack” is still small business territory, both because of scale as well as security concerns.
Big players that don’t have their own datacenters - which includes many tech players - rent private wire cages full of racks, or else “suites” - essentially private rooms. But they’re not typically using companies like Yandex. They use enterprise datacenter providers like Equinix, that require enterprise commitments, and don’t sell to individuals.
No. Some consequences lasted for a few days. It’s all fine now.
yandex.com's search results are currently only showing a single page of 10 results.
Works fine for me ("test" returned 5 pages). But try https://ya.ru too, for example.
I cannot get it to return anything more than 10 results. Might be a design choice. Possibly for mobile. It’s certainly cleaner than the American options that way.
Confusing crawlbots for spambots missed the far bigger story: Russian government has encouraged adoption of digital services and centralization of them into a small number of handpicked "winners", who form the backbone of state surveillance system. This made the modern infrastructure a very attractive target. Ukrainian strikes on the datacenters are comparable to AWS, Google, Stripe and Dropbox going down, with no timeline when the services would be restored. It's been a week and Yandex still shows everything in red: https://status.yandex.cloud/en/dashboard
Just another reminder that false information spreads very fast and will continue to exist for a long time, even if all news provides updates about false information, because such news will not be engaging So, this is a good example of how propaganda works
You know, since gaining access to English Wikipedia and the associated primary/secondary sources, and since watching a lot of edutainment on YouTube, I've come to comprehend just how many lies I've believed in my lifetime. Every headline a half-truth. Every misattributed quote, a white lie. Every factoid, covering something up; misdirecting us from real truth.
Even many things about the Bible, and key theological controversies, have glib formulaic answers from the mouth of our priests. But it turns out that there are deeper truths. It's not that the priests are consciously lying, but they believe these glib answers, and perhaps nobody simply cares to get to the crux of the matter, but just wants a priest to reassure us that we're not wrong?
I'm tired of believing lies. I still read HN's b.s. headlines all day. I still watch a bunch of video news on YouTube. I know, in my heart, that every news story is a half-truth, deliberately concealing enough information to get my engagement or outrage or sympathy. I am addicted to lies and misinformation. I thought I was a truth-seeker, but I wallow in a world of fakes. God help us all.
Wouldn't it be easier to figure out where the bots originate now?
What other providers did they shift to? So we can block them?
Most social media bots are not hosted in datacenters. They are run on racks of mobile phones on regular cellular networks. Using real phones and residential IPs makes them harder to detect. These are often hosted around the world so they can appear as in your country while being remote controlled from Russia.
From the article:
>Our analysis here is a reminder that client ASN traffic volume is an unreliable proxy for interpreting automated bot behavior. Defending against automated threats requires inspecting request behavior, such as identifying shared JA4 fingerprints, monitoring anomalous request rates, and applying WAF rules to sensitive endpoints, regardless of which network the request originates from. Physical infrastructure is static, but malicious automation is built to move.
Of course they're tooting their own horn that you need their advanced features and expertise to block all the bots, but it's approximately true. At the very least it's hard to block purely on ASN without inflicting massive collateral damage on privacy conscious users (ie. using VPN), because baddies can (and do) use the same commercial VPNs.
Two hosts they name in the post are Radist and Selectel
What if they moved to AWS?
Residential proxies
I guess Ukraine has more work to do to cull the trolls.
Would have been helpful if all these cloud corps had multiple ASNs.
I don't understand why they are just starting with the assumption that all of the bot traffic would just go straight out to the internet from their ASNs
If it were a professional operation, the first thing you would do would be to not give away the source of where the traffic is coming from and route the traffic through an intermediary.
Do you think at this point you’re just looking for reasons to believe?
are these "alternate cloud providers" inside Russia?
because Ukraine could offer datacenter-denial as a service
Huh? Surely bot traffic to 'Publishers and Newspapers' falling by 90.3% shows quite clearly that Russian bots were taken offline