4 ms·

we are having occasional lows in page speed performance due to LARGE amounts of bot traffic. full disclosure - we've not really been able to resolve this fully/

by JSeiko 5mo ago

we are having occasional lows in page speed performance due to LARGE amounts of bot traffic. full disclosure - we've not really been able to resolve this fully/well. Let us know if you have a good idea for how to deal with it

TurdF3rguson 5mo ago
CF cache?
lucb1e 5mo ago
I'm only a small-scale sysadmin but the way that I understand the internet is that you send abuse notifications to the IP address block owner and, if it doesn't get resolved, you block. The whois/rdap database reveals which IPs all belong to the same hosting provider or ISP, so you can summarize that all to one list of IP addrs + timestamps per some time period The ISP actually knows which subscriber is on that line, can send them notices, block them, terminate them... loads of things that you simply cannot do because you have no relation to this person. And frankly I wouldn't want to need to have a personal relation with every website that I visit; my ISP can reach me if there is anything relevant to continued use of the internet. From personal experience, when I was a teenager, the ISP cutting our household off after an abuse report was an effective way of stopping what I was doing
- Jolter 5mo ago
  It’s effective against teenagers maybe. Not so much against Amazon, Meta or wherever botnet/crawler is coming out of China these days from up-and-coming AI companies.
  - tonetegeatinst 5mo ago
    I mean you could block entire AS numbers that relate to amazon or big tech datacenters
    - tangledhelix 5mo ago
      wouldn't help, much of the traffic we've observed look closer to ddos patterns - IPs from all over the world, many different networks, each IP makes one request only, doesn't come back. highly distributed, no form of blocking would be effective except maybe captcha or proof of work.
  - lucb1e 5mo ago
    Then block all of Amazon, Meta, or wherever botnet/crawling traffic is coming from that doesn't honor robots.txt, sends DDoS reflection traffic, submits SMTP messages (in large volumes, not just probing) for domains they're not authorized for with SPF, or whatever else applies to the protocol you're using If they can't keep their ranges clean to a reasonable degree, their customers will need to move if they want to access your part of the internet. New sign-ups will always be hard, so some amount of abuse is expected, but if it's the same abuse traffic for weeks after you've notified them, well, it stops being your problem at some point
    - Jolter 5mo ago
      See the other comments in this thread. The perpetrators are unknown and are jumping between residential IPs. Possibly botnets?
      - lucb1e 5mo ago
        Then see my other replies in the thread where I've specifically addressed residential IPs, e.g.: https://news.ycombinator.com/item?id=48163060 https://news.ycombinator.com/item?id=48163060
        Jolter 5mo ago
        This is the post I’m talking about. Make sure you understand how it would not be productive to go after each ISP individually when the traffic is from all of them. https://news.ycombinator.com/item?id=48155512 https://news.ycombinator.com/item?id=48155512
- miki123211 5mo ago
  The problem with this approach is that modern scrapers use hordes of residential proxies and quickly rotate through IP addresses which belong to ASes you get a lot of real traffic from. There's nothing you can do if the ISP won't take any action against the customer.
  - tangledhelix 5mo ago
    Worse than that - even if they would take action, you can't possibly orchestrate filing all of the complaints. It's a drown-in-quicksand problem, you can't fight quicksand one grain at a time.
    - lucb1e 5mo ago
      > you can't possibly orchestrate filing all of the complaints To the ISPs? Each IP range has an abuse email address registered and this is specifically exempt from rate limiting at RIPE's WHOIS server. Not sure how it is in other RIRs but I just happen to know of this policy You can automate the whole thing, provided that you have a reliable way of identifying the undesired traffic which you need anyway for being able to block it by any means. The trouble is in user identification (they'll just use a new IP address from that ISP or hosting provider if you don't tell the provider about the problematic user)
      - tangledhelix 5mo ago
        See what I wrote above (and let me say I am talking about Project Gutenberg and Distributed Proofreaders here, I am one of the admins on both). A large amount of the hassle traffic we've seen is as I wrote above, the IPs come from everywhere and in many cases, each IP makes a single request and doesn't come back. They change user-agent dynamically, etc, to masquerade as regular traffic. They come from residential, cloud/hyperscale, corporate, educational, government, all the networks, on every continent. This is many thousands of "open a ticket with someone" events per hour territory. It's as difficult to fight as DDoS itself for the same reasons (presumably the harvesting parties know that and that's exactly why this approach is used). Others online have been writing about their own experience with the same stuff; it's not unique to PG at all, it's everywhere. Talk to anyone that runs a web server and they'll have these stories...
gropo 5mo ago
Do you host a torrent? I have about 50k of the books, I would have used a torrent of just the txt files if it was prominent.
- gluejar 5mo ago
  we have a tarball of all text files - link posted somewhere here
dimava 5mo ago
If it's purely bot traffic, then Anubis could help You could have seen it on some websites already https://anubis.techaro.lol/ https://anubis.techaro.lol/
- TheDong 5mo ago
  anubis only works against lazy scrapers, and at a cost to your users. I'd prefer people not use it. Bot traffic comes from machines that usually have a lot of idle cpu (since they're largely blocked on network IO as they scrape a bunch of sites in parallel), so they can trivially solve the anubis "proof of work" challenge, save the cookie, and then not solve it again for that site. The only reason scrapers don't solve it is if the developers were too lazy to implement it... and modern scrapers also do, codeberg stopped using anubis because modern scrapers were updated to solve it. The "proof of work" has to be easy or else people on old cell phones couldn't access your site (since an old android phone would start to overheat and throttle trying to solve a challenge that would take a modern server even several seconds), and it also consumes your cell-phone user's batteries, which is a really precious resource for them compared to the idle cpu on a server.
- autoexec 5mo ago
  Please no. I'm a non-bot who gets stopped and turned away all the time by that menace. Anubis doesn't work without JS. One of the things I give duckduckgo a lot of credit for is that while they're quick to interrupt me for a bot check (sometimes multiple times in a span of minutes) they'll let me identify ducks even on the most locked down browsers I use.
- lucb1e 5mo ago
  Just to add to the two negative replies, I find Anubis to be the only system that doesn't ever get in the way. My browsers have Javascript enabled and, so far, it never took more than a fraction of a second to complete the checks Every other system I've run into has constant false positives, e.g. Google captchas will sometimes say I've failed and make me do the hardest level (if it wasn't giving me that already), Cloudflare regularly thinks I'm a bot, Codeberg blocked me before, Github signup captchas used to take ~15 minutes to complete and then still said "well you failed, try again", Github's general rate limiting has false positives (some days I browse a lot, other days little, and on the little days it'll sometimes go "slow down" with no recourse whatsoever, you're just blocked for an indeterminate amount of time), OpenStreetMap blocks my browser at work because I'm using Firefox ESR instead of latest stable and it finds that user agent string to be implausible, whatever the german railway operator uses since a few days is triggering on me constantly, etc., etc., etc. Constant blocks everywhere. With Anubis, my understanding is that you do the proof of work (with whatever implementation you like, it doesn't have to be the Javascript one that they provide) and you can move on without ever doing any task yourself. The power consumption is a shame, but so long as attackers aren't even doing this much, the couple Joules it takes doesn't seem to be an issue Of course, the attackers will evolve, but for now...
jimnotgym 5mo ago
I would love it if you could detect AI scraper bots, and feed them AI generated bs instead of the real books...
- miki123211 5mo ago
  This is very, very, very dangerous. Occasionally, you misclassify a real user as a bot, and then your reputation is ruined forever. The official Polish train schedules website did this recently, feeding incorrect departure and arrival times to IP addresses known for aggressive scraping, without taking CGNAT into account. People... have noticed[1]. [1] (Polish) https://zaufanatrzeciastrona.pl/post/kto-i-dlaczego-losuje-w-polsce-rozklad-jazdy-pkp/ https://zaufanatrzeciastrona.pl/post/kto-i-dlaczego-losuje-w...
- tangledhelix 5mo ago
  Cloudflare sells that as a product, they call it Labyrinth IIRC.
uyzstvqs 5mo ago
How do you currently host everything? Your main web server should not be responsible for hosting content. All books should be hosted on mirrors, and clicking download should automatically select a mirror to download it from. Furthermore: * Make sure that all books are downloadable in bulk as torrents. * Every day, generate a CSV file of all available books and their metadata. Distribute this so that bots and user clients can run queries locally, instead of using your search engine.

by JSeiko 5mo ago

TurdF3rguson 5mo ago

CF cache?

lucb1e 5mo ago

I'm only a small-scale sysadmin but the way that I understand the internet is that you send abuse notifications to the IP address block owner and, if it doesn't get resolved, you block. The whois/rdap database reveals which IPs all belong to the same hosting provider or ISP, so you can summarize that all to one list of IP addrs + timestamps per some time period The ISP actually knows which subscriber is on that line, can send them notices, block them, terminate them... loads of things that you simply cannot do because you have no relation to this person. And frankly I wouldn't want to need to have a personal relation with every website that I visit; my ISP can reach me if there is anything relevant to continued use of the internet. From personal experience, when I was a teenager, the ISP cutting our household off after an abuse report was an effective way of stopping what I was doing

Jolter 5mo ago
It’s effective against teenagers maybe. Not so much against Amazon, Meta or wherever botnet/crawler is coming out of China these days from up-and-coming AI companies.
- tonetegeatinst 5mo ago
  I mean you could block entire AS numbers that relate to amazon or big tech datacenters
  - tangledhelix 5mo ago
    wouldn't help, much of the traffic we've observed look closer to ddos patterns - IPs from all over the world, many different networks, each IP makes one request only, doesn't come back. highly distributed, no form of blocking would be effective except maybe captcha or proof of work.
- lucb1e 5mo ago
  Then block all of Amazon, Meta, or wherever botnet/crawling traffic is coming from that doesn't honor robots.txt, sends DDoS reflection traffic, submits SMTP messages (in large volumes, not just probing) for domains they're not authorized for with SPF, or whatever else applies to the protocol you're using If they can't keep their ranges clean to a reasonable degree, their customers will need to move if they want to access your part of the internet. New sign-ups will always be hard, so some amount of abuse is expected, but if it's the same abuse traffic for weeks after you've notified them, well, it stops being your problem at some point
  - Jolter 5mo ago
    See the other comments in this thread. The perpetrators are unknown and are jumping between residential IPs. Possibly botnets?
    - lucb1e 5mo ago
      Then see my other replies in the thread where I've specifically addressed residential IPs, e.g.: https://news.ycombinator.com/item?id=48163060 https://news.ycombinator.com/item?id=48163060
      - Jolter 5mo ago
        This is the post I’m talking about. Make sure you understand how it would not be productive to go after each ISP individually when the traffic is from all of them. https://news.ycombinator.com/item?id=48155512 https://news.ycombinator.com/item?id=48155512
miki123211 5mo ago
The problem with this approach is that modern scrapers use hordes of residential proxies and quickly rotate through IP addresses which belong to ASes you get a lot of real traffic from. There's nothing you can do if the ISP won't take any action against the customer.
- tangledhelix 5mo ago
  Worse than that - even if they would take action, you can't possibly orchestrate filing all of the complaints. It's a drown-in-quicksand problem, you can't fight quicksand one grain at a time.
  - lucb1e 5mo ago
    > you can't possibly orchestrate filing all of the complaints To the ISPs? Each IP range has an abuse email address registered and this is specifically exempt from rate limiting at RIPE's WHOIS server. Not sure how it is in other RIRs but I just happen to know of this policy You can automate the whole thing, provided that you have a reliable way of identifying the undesired traffic which you need anyway for being able to block it by any means. The trouble is in user identification (they'll just use a new IP address from that ISP or hosting provider if you don't tell the provider about the problematic user)
    - tangledhelix 5mo ago
      See what I wrote above (and let me say I am talking about Project Gutenberg and Distributed Proofreaders here, I am one of the admins on both). A large amount of the hassle traffic we've seen is as I wrote above, the IPs come from everywhere and in many cases, each IP makes a single request and doesn't come back. They change user-agent dynamically, etc, to masquerade as regular traffic. They come from residential, cloud/hyperscale, corporate, educational, government, all the networks, on every continent. This is many thousands of "open a ticket with someone" events per hour territory. It's as difficult to fight as DDoS itself for the same reasons (presumably the harvesting parties know that and that's exactly why this approach is used). Others online have been writing about their own experience with the same stuff; it's not unique to PG at all, it's everywhere. Talk to anyone that runs a web server and they'll have these stories...

gropo 5mo ago

Do you host a torrent? I have about 50k of the books, I would have used a torrent of just the txt files if it was prominent.

gluejar 5mo ago
we have a tarball of all text files - link posted somewhere here

dimava 5mo ago

If it's purely bot traffic, then Anubis could help You could have seen it on some websites already https://anubis.techaro.lol/ https://anubis.techaro.lol/

TheDong 5mo ago
anubis only works against lazy scrapers, and at a cost to your users. I'd prefer people not use it. Bot traffic comes from machines that usually have a lot of idle cpu (since they're largely blocked on network IO as they scrape a bunch of sites in parallel), so they can trivially solve the anubis "proof of work" challenge, save the cookie, and then not solve it again for that site. The only reason scrapers don't solve it is if the developers were too lazy to implement it... and modern scrapers also do, codeberg stopped using anubis because modern scrapers were updated to solve it. The "proof of work" has to be easy or else people on old cell phones couldn't access your site (since an old android phone would start to overheat and throttle trying to solve a challenge that would take a modern server even several seconds), and it also consumes your cell-phone user's batteries, which is a really precious resource for them compared to the idle cpu on a server.
autoexec 5mo ago
Please no. I'm a non-bot who gets stopped and turned away all the time by that menace. Anubis doesn't work without JS. One of the things I give duckduckgo a lot of credit for is that while they're quick to interrupt me for a bot check (sometimes multiple times in a span of minutes) they'll let me identify ducks even on the most locked down browsers I use.
lucb1e 5mo ago
Just to add to the two negative replies, I find Anubis to be the only system that doesn't ever get in the way. My browsers have Javascript enabled and, so far, it never took more than a fraction of a second to complete the checks Every other system I've run into has constant false positives, e.g. Google captchas will sometimes say I've failed and make me do the hardest level (if it wasn't giving me that already), Cloudflare regularly thinks I'm a bot, Codeberg blocked me before, Github signup captchas used to take ~15 minutes to complete and then still said "well you failed, try again", Github's general rate limiting has false positives (some days I browse a lot, other days little, and on the little days it'll sometimes go "slow down" with no recourse whatsoever, you're just blocked for an indeterminate amount of time), OpenStreetMap blocks my browser at work because I'm using Firefox ESR instead of latest stable and it finds that user agent string to be implausible, whatever the german railway operator uses since a few days is triggering on me constantly, etc., etc., etc. Constant blocks everywhere. With Anubis, my understanding is that you do the proof of work (with whatever implementation you like, it doesn't have to be the Javascript one that they provide) and you can move on without ever doing any task yourself. The power consumption is a shame, but so long as attackers aren't even doing this much, the couple Joules it takes doesn't seem to be an issue Of course, the attackers will evolve, but for now...

jimnotgym 5mo ago

I would love it if you could detect AI scraper bots, and feed them AI generated bs instead of the real books...

miki123211 5mo ago
This is very, very, very dangerous. Occasionally, you misclassify a real user as a bot, and then your reputation is ruined forever. The official Polish train schedules website did this recently, feeding incorrect departure and arrival times to IP addresses known for aggressive scraping, without taking CGNAT into account. People... have noticed[1]. [1] (Polish) https://zaufanatrzeciastrona.pl/post/kto-i-dlaczego-losuje-w-polsce-rozklad-jazdy-pkp/ https://zaufanatrzeciastrona.pl/post/kto-i-dlaczego-losuje-w...
tangledhelix 5mo ago
Cloudflare sells that as a product, they call it Labyrinth IIRC.

uyzstvqs 5mo ago

How do you currently host everything? Your main web server should not be responsible for hosting content. All books should be hosted on mirrors, and clicking download should automatically select a mirror to download it from. Furthermore: * Make sure that all books are downloadable in bulk as torrents. * Every day, generate a CSV file of all available books and their metadata. Distribute this so that bots and user clients can run queries locally, instead of using your search engine.