New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
Life happen. You are always welcome to put some effort on a project, open source it and keep working on it for free. Is normal for open source projects to eventually commercialize.
I have no idea why you bothered quoting my entire post- or, indeed, responding to me when I was just illustrating how you can get away with not following the GPL license (legally).
The time-tested and true means is usually to have at least two forks on a project/product, and to do work on the commercial product with security backports and possible features added to the GPL product later (even if it's trvial to do both unless you're horribly sloppy).
Matt has shown direct intent to make it difficult for anyone to produce non-official binaries of Caddy. That is a very Theo thing to do- even though Theo has calmed down significantly since the late 90s.
This is indeed a good strategy. Virtuozzo/OpenVZ does this well.
I'm sure someone - a student maybe or some dedicated devs - will fork Caddy and continue development.
Flaming someone on Twitter who wants to abet free-riding is not direct intent to sabotage non-official binaries. If you are a struggling artist, how happy are you when your fans just download your music for nothing?
While expensive for licensing, totally understand that part.
But using HTTP header as a means for it isn't my cup of tea when there are known performance overhead hit as you add more headers to Caddy. Though you can build from source yourself to get around the HTTP sponsor header requirement https://caddy.community/t/caddy-commercial-sponsor-header-clarification/2716
But right now building from source is sort of broken too https://github.com/mholt/caddy/issues/1843.
I think most personal usage folks using Caddy wouldn't care about that as they'd just use free version/personal with HTTP sponsor header intact anyway not knowing or caring about the performance hit as you add more HTTP headers.
For HTTP/2 HTTPS loads, Caddy is ~1/3rd the performance of Nginx. So you'd need 3-4x Caddy servers to match the performance of 1x Nginx server for HTTP/2 based HTTPS. So for Commercial licensing you'd need 5/server license at discounted US$250/month which reverts to 4x250 = US$1,000 month once introductory licensing ends. That equates to $3,000/yr discounted or $12,000/yr for 5 commercial licenses if you only need 4 ?
Compare Nginx commercial licensing for 1-4 servers https://www.nginx.com/products/pricing/
If I need 4x Caddy servers to match 1x Nginx servers performance the comparative cost is:
If I need 8x Caddy servers to match 2x Nginx servers performance the comparative cost is:
From financial and performance perspective, doesn't make sense unfortunately.
That's not at all what I have illustrated in this very thread as being hostile towards other devs, and most artists deserve to starve. Please look up.
This is why we like you. Smart.
cheers - I'm a logics guy
There's better ways to do it like Nginx vs Nginx Plus, their commercial version just adds more value with advanced options and features compared to open source. Though I guess Caddy are trying to replicating that with the Private Plugin Hosting etc.
Would like to see the performance hit from one small header. Suspect it would be quite acceptable, if one preferred the web server for any reason.
Granted, would like to see the benefits as well since basically no one will see it.
According to @eva2000, it's already -66% of an nginx service (I have not installed nor tested this numeric). As you know, other than adding trivial overhead which is possibly 1k (if ever reaching that), is easy to just add to ones' existing traffic.
The single "header" is NOT what everyone is annoyed by, for what it's worth. You can continue to pretend that's why everyone's up in arms- but it isn't.
The changing of the licensing is just pure greed. The fact that the new license denies the abilities of a commercial entity to use this product at all without a license is going to alienate the people who actually like/want to use this product.
Their pricing tier does not make any sense- especially for a product that isn't proven to be better than anything else on the market (and is evidently much slower).
Matt's trying to stall/block other developers from being able to build from a fork of the existing source tree tends to be what has annoyed most developers that I am aware of.
And now one of the sites he's included in the "This free server is supported by:" header, Minio, has asked him to remove their company from it. https://github.com/mholt/caddy/pull/1866 (abperiasamy)
I wonder how sponsors will feel if the server was serving some kind of illegal content. "This server is supported by: "
Really? So what is even the point to bother with it in the first place.
In any case, I suppose as long as it is needed by someone, there will be a fully Free and OSS fork (Wedge now), folding in all the useful changes and updates, while removing all the annoyances and providing official unencumbered binaries. With time most users may even migrate to that. This all is just FOSS working as intended.
misunderstood Caddy even before HTTP header performance hits, already runs at ~33% of Nginx performance. Adding additional HTTP headers will lower that somewhat more ~5-10% more.
All benchmarks are at https://community.centminmod.com/threads/caddy-http-2-server-benchmarks.5170/
That's quite hard to believe from a purely programmatic standpoint.
As an aside, interesting to note that with Nginx you cannot adapt the 'Server:' HTTP header. It's a hardwired thing.
Why would anyone want to use Caddy, let alone pay for it?
issue happens on both h2o and caddy web servers from my tests but not nginx
https://forum.caddyserver.com/t/any-performance-overhead-as-you-add-more-headers-under-http-2/403/3
h2o reported issue https://github.com/h2o/h2o/issues/240
FYI, Caddy's new http.cache proxy cache plugin helps with raising Caddy performance but still behind Nginx. See benchmarks I did at https://caddy.community/t/announce-new-http-cache-plugin/2429/10
Is quite hard to beat serverpilot.io stack which combines apache with nginx and a few other relevant technologies. Heck is already hard enough to beat Nginx alone... Why would anyone code a new web servers, make it open source then try to monetize it when is a lot slower than the alternatives?
Reason: Hipsters.
@ricardo You can change the server header with the "Extra Headers" module for NGINX.
Becouse installation and configuration takes a couple of seconds? On top of that it handles your certificates for you. Ideal for dev stacks. Would I be prepared to pay for it? No. With their pricing it would make more sense to fork it myself and hire a programmer to maintain it.
That's what I thought, too, but evidently "new" programming is something we've never even considered since they can manage concurrence and automatic garbage collection, but appending a small bit of text causes the whole thing to crash to a grinding halt.
Ahh yes. I remember coming across this on an older version of NGINX but now you can use this module. I was trying to avoid the fingerprint of showing the same backend proxied content. Their original reasoning for hard coding it was for metrics & branding IIRC.
Yeah, I mean if the test was a static page of say, a hundred bytes or so, I could see why the additional writes & tiny overhead of a hook would result in that kind of difference.
Build a tentacle porn site using caddy.
long discussion on this at https://news.ycombinator.com/item?id=15237923 too
imagine they had 100 sponsor
It'd take 4 minutes to render a page!
Seems like they might have lost minio as a sponsor https://github.com/mholt/caddy/pull/1866
https://caddyserver.com/sponsor
So what's the LET summation... Guy working on his lonesome gets shat on by freetards? Demonic bait and switch spitting in the face of open source? Greedy bastard cashing in?
I'll continue using Nginx but it's interesting to see the reaction. I don't know enough to see how much has been contributed to the repo by the community
too much.