You said that your test case peaks at 600Mbit/sec.

Your first step should be to bisect the problem, to see if you're limited by your hardware+OS or your test + nginx configuration. Easiest way is to install solarflare's free network test utility from the support section of their website.

After that, to dig further into web specific factors it can be worth install the (large) Tech Empower web framework test bed. This is a set of five or six micro benchmarks implemented in over 80 web frameworks. You can compare your results with their published results.

Peter

Sent from my iPhone

> On Jan 7, 2017, at 7:00 AM, nginx-request@nginx.org wrote:
>
> Send nginx mailing list submissions to
> nginx@nginx.org
>
> To subscribe or unsubscribe via the World Wide Web, visit
> http://mailman.nginx.org/mailman/listinfo/nginx
> or, via email, send a message with subject or body 'help' to
> nginx-request@nginx.org
>
> You can reach the person managing the list at
> nginx-owner@nginx.org
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of nginx digest..."
>
>
> Today's Topics:
>
> 1. Performance test caps at 600 Mbit/s (MrFastDie)
> 2. upstream prematurely closed connection (Thierry)
> 3. Re: Performance test caps at 600 Mbit/s (CJ Ess)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 06 Jan 2017 10:24:06 -0500
> From: "MrFastDie" <nginx-forum@forum.nginx.org>
> To: nginx@nginx.org
> Subject: Performance test caps at 600 Mbit/s
> Message-ID:
> <c38ed8477b83d9b761f3bb69eb475244.NginxMailingListEnglish@forum.nginx.org>
>
> Content-Type: text/plain; charset=UTF-8
>
> Hello,
>
> the last days I played a little with the NGINX settings and the tcp stack to
> test the best performance. I used direct connection between my testing
> machine and my server using a cat5e cable.
> My nginx.conf can be found at pastebin: http://pastebin.com/rRAEwvNc
> My sysctl.conf also provides lot of changes: http://pastebin.com/KmPjEnHN
> My /etc/security/limits.conf provides soft nofile at 1048576 and hard nofile
> at the same value.
>
> I'm not using any kind of firewall, all logs are disabled, neither the cpu
> nor the ram is at it's limit, the i/o load seems fine to me too, but the
> internet speed caps round about 600 Mbit/s. This test was made with 2000
> connections per time and a 1MB file.
>
> The server contains an Intel(R) Pentium(R) Dual CPU E2220 @ 2.40GHz and
> has 3.5G of Ram.
> My local computer contains an Intel(R) Core(TM) i7-4710HQ CPU @ 2.50GHz and
> has 16G of Ram.
>
> The limits and tcp changes were also made on my local computer.
>
> Is there someone who can help me to deal with this problem?
>
> Posted at Nginx Forum: https://forum.nginx.org/read.php?2,271846,271846#msg-271846
>
>
>
> ------------------------------
>
> Message: 2
> Date: Fri, 6 Jan 2017 17:28:59 +0200
> From: Thierry <lenaigst@maelenn.org>
> To: nginx@nginx.org
> Subject: upstream prematurely closed connection
> Message-ID: <1986304318.20170106172859@maelenn.org>
> Content-Type: text/plain; charset=us-ascii
>
> Dear all,
>
> The error message:
> 2017/01/05 18:28:49 [error] 32633#32633: *3 upstream prematurely closed connection while reading response header from upstream, client: xxx.xxx.xxx.xxx, server: server_hostname.domain.ltd, request: "POST /SOGo/connect HTTP/2.0", upstream: "https://ip_server:port_number/SOGo/connect", host: "server_hostname.domain.ltd:port_number", referrer: "https://server_hostname.domain.ltd:port_number/SOGo/"
> 2017/01/05 18:28:49 [error] 32633#32633: *3 upstream prematurely closed connection while reading response header from upstream, client: xxx.xxx.xxx.xxx, server: server_hostname.domain.ltd, request: "GET /SOGo.woa/WebServerResources/busy.gif HTTP/2.0", upstream: "https://ip_server:port_number/SOGo.woa/WebServerResources/busy.gif", host: "server_hostname.domain.ltd:port_number", referrer: "https://server_hostname.domain.ltd:port_number/SOGo/"
>
> I have a front-end reverse proxy using Nginx for my web server and
> for the web interface of my email server using Sogo.
>
> From my firewall:
> - If I NAT the "Sogo" port to my reverse-proxy I do
> have these error and a 502 Bad Gateway message.
> - If I bypass my reverse-proxy and NAT directly to my email server,
> this is working.
>
> On my reverse-proxy server, I am using Nginx version 1.9.10 (debian 8)
> On my email server, I am using Nginx version 1.6.2 (debian 8)
>
> If you need more information, please ask ... I do not understand
> this error myself.
>
> ps: it is the second time I am sending this email (?)
>
>
> --
> Cordialement,
> Thierry e-mail : lenaigst@maelenn.org
> PGP Key: 0xB7E3B9CD
>
>
>
> ------------------------------
>
> Message: 3
> Date: Fri, 6 Jan 2017 11:25:47 -0500
> From: CJ Ess <zxcvbn4038@gmail.com>
> To: nginx@nginx.org
> Subject: Re: Performance test caps at 600 Mbit/s
> Message-ID:
> <CA+p3EyyTC+iqS-4uU1-Gfs8JX_mMTNJK3bo_ZX-eZJ16mA6_8A@mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
>
> Which OS? What NIC? You also have to consider the traffic source, is it
> known capable of of saturating the NIC on your server?
>
> On Fri, Jan 6, 2017 at 10:24 AM, MrFastDie <nginx-forum@forum.nginx.org>
> wrote:
>
>> Hello,
>>
>> the last days I played a little with the NGINX settings and the tcp stack
>> to
>> test the best performance. I used direct connection between my testing
>> machine and my server using a cat5e cable.
>> My nginx.conf can be found at pastebin: http://pastebin.com/rRAEwvNc
>> My sysctl.conf also provides lot of changes: http://pastebin.com/KmPjEnHN
>> My /etc/security/limits.conf provides soft nofile at 1048576 and hard
>> nofile
>> at the same value.
>>
>> I'm not using any kind of firewall, all logs are disabled, neither the cpu
>> nor the ram is at it's limit, the i/o load seems fine to me too, but the
>> internet speed caps round about 600 Mbit/s. This test was made with 2000
>> connections per time and a 1MB file.
>>
>> The server contains an Intel(R) Pentium(R) Dual CPU E2220 @ 2.40GHz and
>> has 3.5G of Ram.
>> My local computer contains an Intel(R) Core(TM) i7-4710HQ CPU @ 2.50GHz and
>> has 16G of Ram.
>>
>> The limits and tcp changes were also made on my local computer.
>>
>> Is there someone who can help me to deal with this problem?
>>
>> Posted at Nginx Forum: https://forum.nginx.org/read.
>> php?2,271846,271846#msg-271846
>>
>> _______________________________________________
>> nginx mailing list
>> nginx@nginx.org
>> http://mailman.nginx.org/mailman/listinfo/nginx
>>
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: http://mailman.nginx.org/pipermail/nginx/attachments/20170106/148b3196/attachment-0001.html
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> nginx mailing list
> nginx@nginx.org
> http://mailman.nginx.org/mailman/listinfo/nginx
>
> ------------------------------
>
> End of nginx Digest, Vol 87, Issue 7
> ************************************
_______________________________________________
nginx mailing list
nginx@nginx.org
http://mailman.nginx.org/mailman/listinfo/nginx