Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

OpenFreeMap survived 100k requests per second

blog.hyperknot.com

Post reply on HN
Print view View original

121–130 of 140 posts

Re: OpenFreeMap survived 100k requests per second

#121

Post by johnisgood »

zahlman wrote: post #115
Starlevel004 wrote: post #57

> I believe what is happening is that those images are being drawn by some script-kiddies. If I understand correctly, the website limited everyone to 1 pixel per 30 seconds, so I guess everyone was just scripting Puppeteer/Chromium to start a new browser, click a pixel, and close the browser, possibly with IP address rotation, but maybe that wasn't even needed. I think you perhaps underestimate just how big of a thin…

> I think you perhaps underestimate just how big of a thing this became basically overnight. I mentioned a drawing over my house to a few people and literally everyone instantly knew what I meant without even saying the website. On the other hand, this is the first I've heard of this thing.

I have known about this kind of pixel drawing but it was on empty canvas.

Re: OpenFreeMap survived 100k requests per second

#122

Post by drewda »

Ericson2314 wrote: post #80

Oh wow, TIL there is finally a simple way to actually view OpenStreetMap! Gosh, that's overdue. Glad it's done though!

The OSM Foundation has been serving raster tiles for years and years (that's what's visible by default on the slippy map at www.openstreetmap.org): https://wiki.openstreetmap.org/wiki/OpenStreetMap_Carto

After on and off experimentation by various contributors, OSMF just released vector tiles as well: https://operations.osmfoundation.org/policies/vector/

Re: OpenFreeMap survived 100k requests per second

#123

Post by johnisgood »

Aeolun wrote: post #101
toast0 wrote: post #54

Earlier quoted context omitted.

I'm pretty sure your open file cache is way too large. If you're doing 1k/sec, and you cache file descriptors for 60 minutes, assuming those are all unique, that's asking for 3 million FDs to be cached, when you've only got 1 million available. I've never used nginx or open_file_cache[1], but I would tune it way down and see if you even notice a difference in performance in normal operation. Maybe 10k files, 60s time…

If you do 200Mbps on a hetzner server after cloudflare caching, you are going to run out of traffic pretty rapidly. The limit is 20TB / month (which you’d reach in roughly 9 days).

One would think services like these do not have to rely on online services and have their own rack of servers. Or is this so alien these days?

Re: OpenFreeMap survived 100k requests per second

#124

Post by indrora »

Starlevel004 wrote: post #57

> I believe what is happening is that those images are being drawn by some script-kiddies. If I understand correctly, the website limited everyone to 1 pixel per 30 seconds, so I guess everyone was just scripting Puppeteer/Chromium to start a new browser, click a pixel, and close the browser, possibly with IP address rotation, but maybe that wasn't even needed. I think you perhaps underestimate just how big of a thin…

It's also way more than 1px/30s -- Its like 20px/30s and you have a "tank" of them, which you can expand to however big you want.

Placing pixels gives you points, which you can turn into more pixels or a bigger bag of pixels over time. I've seen people who have done enough pixel pushing that they get 3-4K pixels at a time.

Re: OpenFreeMap survived 100k requests per second

#125

Post by MaKey »

Aeolun wrote: post #101
toast0 wrote: post #54

Earlier quoted context omitted.

I'm pretty sure your open file cache is way too large. If you're doing 1k/sec, and you cache file descriptors for 60 minutes, assuming those are all unique, that's asking for 3 million FDs to be cached, when you've only got 1 million available. I've never used nginx or open_file_cache[1], but I would tune it way down and see if you even notice a difference in performance in normal operation. Maybe 10k files, 60s time…

If you do 200Mbps on a hetzner server after cloudflare caching, you are going to run out of traffic pretty rapidly. The limit is 20TB / month (which you’d reach in roughly 9 days).

Small addition: That limit applies to Hetzner Cloud servers, their dedicated servers have unlimited traffic.

Re: OpenFreeMap survived 100k requests per second

#126

Post by zamadatix »

yifanl wrote: post #118
Starlevel004 wrote: post #57

> I believe what is happening is that those images are being drawn by some script-kiddies. If I understand correctly, the website limited everyone to 1 pixel per 30 seconds, so I guess everyone was just scripting Puppeteer/Chromium to start a new browser, click a pixel, and close the browser, possibly with IP address rotation, but maybe that wasn't even needed. I think you perhaps underestimate just how big of a thin…

They have the user count from the dev, 2 million daily users shouldn't be generating billions of requests unless a good portion of them are botting.

Why not? This is tile requests right, not login requests or something, so shouldn't a single user be expected to consume a few thousand zipping around the map while looking at drawings overlaid?

I'm sure there is some botting, it's basically guaranteed, but I wouldn't be surprised if nearly half the traffic was "legitimate". The bots don't normally need to reload (or even load) the map tiles anyways.

Re: OpenFreeMap survived 100k requests per second

#127

Post by zamadatix »

Aurornis wrote: post #116
ivanjermakov wrote: post #94

> Nice idea, interesting project, next time please contact me before. I understand that my popular service might bring your less popular one to the halt, but please configure it on your end so I know _programmatically_ what its capabilities are. I host no API without rate-limiting. Additionally, clearly listing usage limits might be a good idea.

> I understand that my popular service might bring your less popular one to the halt, but please configure it on your end so I know _programmatically_ what its capabilities are. Quite entitled expectations for someone using a free and open service to underpin their project. The requests were coming from distributed clients, not a central API gateway that could respond to rate limiting requests > I host no API without…

> The requests were coming from distributed clients, not a central API gateway that could respond to rate limiting requests

The block was done based on URL origin rather than client/token, why wouldn't a rate limiter solution consider the same? For this case (a site which uses the API) it would work perfectly fine. Especially since the bots don't even care about the information from this API so non-site based bots aren't even going to bother to pull the OpenFreeMap tiles.

Re: OpenFreeMap survived 100k requests per second

#128

Post by Aeolun »

cyberpunk wrote: post #109
Aeolun wrote: post #98

Earlier quoted context omitted.

> They’re unreliable as soon as you have to deal with their support who have the technical knowledge of a brick. Since I never have to, that’s perfect isn’t it? If you need support from Hetzner you are using the wrong host.

So they've never decided to blackhole your prod traffic 2 days after migrating to them? lucky.

I don’t think I’ve ever migrated anywhere and not kep the previous thing running for a month unless I wasn’t worried about exactly that thing happening (to be fair, not prod traffic blackhole, but similar effect)

Re: OpenFreeMap survived 100k requests per second

#129

Post by Aeolun »

MaKey wrote: post #125
Aeolun wrote: post #101

Earlier quoted context omitted.

If you do 200Mbps on a hetzner server after cloudflare caching, you are going to run out of traffic pretty rapidly. The limit is 20TB / month (which you’d reach in roughly 9 days).

Small addition: That limit applies to Hetzner Cloud servers, their dedicated servers have unlimited traffic.

Depends on your connection I think. Mine do 1Gbit/sec but have a 20TB limit. The 100Mbit ones are unlimited (last I checked)

Re: OpenFreeMap survived 100k requests per second

#130

Post by ranger_danger »

Aeolun wrote: post #98
cyberpunk wrote: post #95

Earlier quoted context omitted.

They’re unreliable as soon as you have to deal with their support who have the technical knowledge of a brick. And as soon as you have to do ant business / deal with the german side of the business expect everything to slow down to 2 weeks for response which will still be incorrect. They are simply not worth the hassle. Go with a competent host.

> They’re unreliable as soon as you have to deal with their support who have the technical knowledge of a brick. Since I never have to, that’s perfect isn’t it? If you need support from Hetzner you are using the wrong host.

Simply running "ipfs daemon" is enough to get hetzner to threaten canceling your service.

They send nasty letters with pcap dumps of your "hacking" attempts and it doesn't matter how wrong you think they are.

Post reply on HN