Live data from Hacker News

OpenFreeMap survived 100k requests per second

blog.hyperknot.com

101–110 of 140 posts

Re: OpenFreeMap survived 100k requests per second

#101
post #54

Earlier quoted context omitted.

I've just written my question to the nginx community forum, after a lengthy debugging session with multiple LLMs. Right now, I believe it was the combination of multi_accept + open_file_cache > worker_rlimit_nofile. https://community.nginx.org/t/too-many-open-files-at-1000-re... Also, the servers were doing 200 Mbps, so I couldn't have kept up _much_ longer, no matter the limits.

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).

Re: OpenFreeMap survived 100k requests per second

#102
post #43

Is it always/only 'laziness' (derogatory, i know) when caching isn't implemented by a site like wplace.live ? Why wouldn't they save openfreemap all the traffic when a caching server on their side presumably could serve tiles almost as fast or faster than openfreemap?

Why should they when openfreemap is behind a CDN and their home page says things like: > Using our public instance is completely free: there are no limits on the number of map views or requests. There’s no registration, no user database, no API keys, and no cookies. We aim to cover the running costs of our public instance through donations. > Is commercial usage allowed? > Yes. IMHO, reading this and then just using…

I think, when you read that, you should be reassured that nobody is going to suddenly tell you to pay, and then still implement caching on your own side to preserve the free offering for everyone else.

Seriously, whose first thought on reading that is “oh great, I can exploit this”.

Re: OpenFreeMap survived 100k requests per second

#103
post #73

Earlier quoted context omitted.

I'm sure the server hardware is not a problem. The full dataset is 150 GB and the server has 64 GB RAM, most of which will be never requested. So I'm sure that the used tiles would actually get served from OS cache. If not, it's on a RAID 0 NVME SSD, connected locally. What I've been referring to is the fact that even unlimited 1 Gbps connections can be quite expensive, now try to find a 2x40 gig connection for a rea…

Well, “bandwidth is expensive” is a true claim, but it's also a very different claim from “a [normal] caching server couldn't handle 56 Gbit/sec”…?

56 Gbit/sec costs you about €590/day even on Hetzner.

Re: OpenFreeMap survived 100k requests per second

#104
post #88

Since cloudflare is already sponsoring it, I do wonder how much of this type of service can be implemented all on cloudflare. Their stack could be great for tile serving.

I’m also surprised to see nginx and hetzner in this project. Why not entirely Cloudflare: workers, R2, and cache

You can get cheap dedicated server on Hetzner with unlimited bandwidth, would the cost be similar with CF?

Re: OpenFreeMap survived 100k requests per second

#105
post #78

Earlier quoted context omitted.

IME Hetzner's not unreliable. I don't think you could serve 100k requests per second on a single VPS though. (And with dedicated, you're on the hook for redundancy yourself, same as any dedicated.)

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.

Ideally, we'd all be using very good and competent hosting companies, but they're not ideal for free/low revenue projects. It's better to have some down time than having to shut down the service because you can't afford to keep it running.

I think that in both cases here (OpenFreeMap and wplace), Hetzner/OVH/Scaleway is the way to go. Depending on what we're doing, the cost savings can even allow us to have redundancy at another cheap provider just in case something goes wrong.

Re: OpenFreeMap survived 100k requests per second

#107
post #102
post #43

Earlier quoted context omitted.

Why should they when openfreemap is behind a CDN and their home page says things like: > Using our public instance is completely free: there are no limits on the number of map views or requests. There’s no registration, no user database, no API keys, and no cookies. We aim to cover the running costs of our public instance through donations. > Is commercial usage allowed? > Yes. IMHO, reading this and then just using…

I think, when you read that, you should be reassured that nobody is going to suddenly tell you to pay, and then still implement caching on your own side to preserve the free offering for everyone else. Seriously, whose first thought on reading that is “oh great, I can exploit this”.

You don't need to be thinking "I can't exploit this" when you can just stop thinking about it.

Re: OpenFreeMap survived 100k requests per second

#109
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. 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.

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