So this one datacenter in Russia got bombed by Ukraine and caused a Yandex outage. And I have seen people posting data-backed analyses showing this has resulted in a drop in net traffic from certain scraper bot networks; and I've seen people making *jokes* that this will show up in drops of coordinated inauthentic activity on websites such as "X" [Twitter]. But I haven't seen any attempts to analyze with hard data whether social media activity changed during this particular data center outage.
1 Boost
0 Likes
@mcc
This is from Dave Troy. I have no idea how good he is at his job, but he seems committed to getting at the data
https://russianbots.america2.news
[Thanks for this thread. I, too, would like a data-driven answer]
@mcc There has been a similar incident in 2024 and this think tank here did collect some data on this but I would say this is still very thin.
https://www.bruegel.org/first-glance/russian-internet-outage-offers-clues-about-online-goblin-army
@mcc
This sort of monitoring is exactly what Twitter's API was great for, before Musk put the whole damn thing behind a paywall.
Meta/TikTok have fought researchers on this kind of stuff from the jump.
@mcc Here you go. Here's a graph of bot activity from Yandex, showing a steep drop-off.
@mcc The image that's been circulating is attributed to "Cloudflare" in this article. https://www.tomshardware.com/tech-industry/data-centers/russian-disinformation-bot-traffic-nearly-eliminated-overnight-after-ukraine-drone-strikes-cripple-multiple-of-russias-google-data-centers-ukraine-intensifies-bombing-of-yandex-data-centers-in-response-to-moscows-strikes-on-kyiv-data-centers You can find a similar chart here: https://radar.cloudflare.com/bots/as13238 That's (as far as I can tell), "bots from Yandex," which I'm sure have decreased recently. But I don't think it's "percentage of bots overall."
@CraigStuntz Thanks! However "bots" as measured by Cloudflare is surely different from "bots" referring to organized inauthentic activity on social networks? Like Cloudflare "bots" probably refer to something like AI scrapers. (Or in the case of Yandex's own bots rather than their cloud service customers, search engine scrapers?)
@mcc Here's someone else looking at numbers from a different source and concluding that rumors of the death of Russian social media influence operations may be overstated: https://bsky.app/profile/katestarbird.bsky.social/post/3mxrjoq7cqs2b
@CraigStuntz Thanks! I also saw this (sorry, Bad Place link) https://x.com/ValdemarOsted/status/2109742982568313237
@mcc No idea, sorry. I didn't dig deep enough to provide any insight on that question.
![Mre. Dartigen [maker mode]](https://mediacdn.aus.social/accounts/avatars/108/195/770/473/016/150/original/d5b937db377b2251.gif "Mre. Dartigen [maker mode]")
@mcc https://mastodon.nz/@KiwiEV/117424302046811118 has a graph of HTTP requests, but I'm not sure to/from what or how it was confirmed to be bot traffic. Might be worth asking OP.
@mcc I've read that the people with the pretty graphs are measuring BGP prefix advertisements, which probably *would* decrease if the routers got blown to smithereens. But I'm not sure.
@theorangetheme That makes sense, but… is that… … sinister? Presumably BGP prefix advertisements would measure how much Russians are using the Internet, not state-sponsored information influence operations.
i always assume traffic somewhere is potentially being used to push a narrative or a psyop. either way for me, data centers are prime targets for disruption because they can potentially help minimize DDoS and counter intelligence.
either way its more direct intervention against russia than anything the west is doing right now.
@cyberquixote @theorangetheme Currently I am not attempting to judge morality or purpose of any action. I am attempting to understand what is happening in the world around me
@mcc Data sources are hard to come by, most of the social media sites can be pretty tight lipped about their traffic. Experiment design would consist of creating some number of accounts to "follow" a representative sample of "bot" accounts and measure posting activity over time. I know the disinfo people do something like this to observe disinfo trends.
@mcc I was thinking that. I'd love to see someone see if there are changes in posting patterns or if some accounts coincidentally stop and restart posting on various websites as connections are re-established.
@mcc I think this is the third datacenter they've hit in as many days
@cthos I think that's true yes. I only started to hear people talk about service disruptions when the third datacenter got hit, so either Yandex had localized all of their operations in the one datacenter, or else there was some sort of failover mechanism that could handle the loss of two datacenters but not three.
@mcc i have seen people (okay, one person) who was trying to *get* actual data, at least
@ratsnakegames i wish them luck
I did see this literally right after this post
@The4thCircle @mcc perhaps unsurprisingly, this graph is entirely miscontextualised.
It's specifically measuring general traffic from Yandex, which is mostly crawlers because Yandex is a search engine
It's specifically measuring general traffic from Yandex, which is mostly crawlers because Yandex is a search engine
@The4thCircle I did see that, but here's the thing: First off there was no cite or explanation of how the data was measured that I found; second off I don't even speak German.
@mcc @The4thCircle i can understand some german, and i would read the title as "daily activity above three months", the x label as "Date" and the y label as "Activity per day".
The title sounds very awkward to me, I would interpret "über 3 Monate" as "greater than three months", while english "for three months" would become "seit 3 Monaten". I'm not particularly good at german though.
@artemist@mildlyfunctional.gay @mcc@mastodon.social @The4thCircle@mastodon.gamedev.place über is very analogous to "over" in English; I'd interpret it as "over [the last] 3 months"
@artemist @The4thCircle Hm. But daily activity of what, measured by who?
@mcc It’s probably a bit too early to tell, but I do sure hope some people are looking into this.
@Sylvhem It seems like you'd need to move pretty quick on gathering data because even with the relatively isolated nature of Russia's commercial internet infrastructure surely migrating to alternate providers can happen on some fairly fast timetable.
@mcc@mastodon.social
I have an idea, gona check a thing