Yandex outage bot activity likely shows its own crawlers

Drones hit a Yandex data center in Sasovo, in the Ryazan region, overnight into October 8, then one in Kaluga on October 9. On October 11, Yandex said its site in Vladimir had been damaged by drones and taken offline. Ukraine hasn’t officially claimed the strikes, though Yandex has blamed Ukraine for the Kaluga attack. They followed Russian strikes on Ukrainian data centers since September.

The outages were wide. Yandex’s voice assistant, music and cloud services went down, along with banking, retail, delivery and transport apps that rely on its servers. Disruptions reached users in Russia and several neighboring countries. Most Yandex systems were still offline after the third strike, including the Yandex Pay payment service, which left users unable to move money, check balances or buy things. Yandex Fintech head Dima Pirogov said the company couldn’t yet give a precise restoration timeline.

The chart came from Cloudflare Radar, a public dashboard run by the internet security company Cloudflare, which sees a large share of the web’s traffic. It tracks automated requests coming from Yandex’s network.

The drop was dramatic. Requests began falling around October 8 and hit near zero early on October 11. At one point that day they stood at 0.3% of the chart’s maximum, compared with 88.2% at the same hour a week earlier. A separate reading put them at 0.4% of the previous two weeks’ peak.

That’s what got shared. Posts called it Russian bot traffic, and some added that bots had once written an estimated 21% to 29% of US Twitter content, with the US midterms only weeks away. Ukrainian President Volodymyr Zelensky also said social media had seemed less hostile after the attacks.

To those implying that Kremlin-controlled social media bot/troll farms became silent after Ukraine hit a Yandex data center, posting this CloudFare chart (https://t.co/yMPllx3TnM):

Hate to disappoint you all ⤵️ pic.twitter.com/mQNKtHd2Bf

— bot blocker | блокировщик ботов (@antibot4navalny) October 11, 2026

Here’s the catch. In computing, a bot is just an automated program. Some are malicious, such as tools that break into accounts or send spam. But many are ordinary and useful, like search engine crawlers that visit websites to index them, or monitoring tools.

Antibot4navalny, an independent project that has tracked Russian troll farms for years, said in a thread on X that the chart measures requests from Yandex’s own automated web crawlers. It has nothing to do with fake accounts posting on social media. Cloudflare’s numbers back this up in two ways:

• Bots are almost all of Yandex’s traffic anyway. Over seven days, bots made up 99.6% of requests from the network. A search company’s servers are mostly machines talking to machines.
• Russian bot traffic overall barely moved. Cloudflare data showed automated traffic from Russia as a whole stayed roughly the same around the strikes. Only Yandex’s slice collapsed.

So the chart shows that Yandex’s servers stopped working. It can’t separate fewer bad bots from fewer working computers. With much of Yandex offline, automated requests would fall whether or not any influence campaign ran there.

Other graphics joined the original. Czech member of the European Parliament Tomáš Zdechovský shared one claiming online activity around Germany’s far-right AfD party fell sharply after the strikes.

Computer scientist Raphael Wimmer said he suspects that chart was made by AI. He pointed to inconsistencies in its timeline, including a data point logged for October 15, a date that hadn’t happened yet. The image seems to have first appeared on Threads, its data source was never established, and the original didn’t mention the AfD at all. A German tech outlet also noted that activity around other German parties dipped that weekend too.

Similar claims spread about accounts backing President Donald Trump, Britain’s Reform UK and political pages in Alberta, Canada. In each case, the screenshots showed views on individual posts, which rise and fall for many reasons, and didn’t prove any link to foreign bot networks.

None of this proves Russian influence campaigns don’t exist. They do, and researchers track them closely. It means this particular chart can’t tell you whether they were hit, and the timing alone isn’t evidence either way.

Before you share the next viral graph, a few quick checks help:

• Find the source. A screenshot with no link to the original data is a red flag.
• Ask what’s measured. Web crawler requests, post views and fake accounts are three different things.
• Look at the dates. Data points from the future are a giveaway that a chart was made up.
• Check for other causes. A dip on a weekend, or across every party, isn’t a smoking gun.

Whether the strikes dented any influence operations is still an open question. The best evidence so far shows a big outage at Yandex, not a sudden silence from Russia’s troll farms.