how to disable upstream statistics collection?

[k0st1an]

how to disable upstream statistics collection?

I want to disable statistics collection for upstream servers to reduce CPU load. How can I do this?

Current config

load_module modules/ngx_http_vhost_traffic_status_module.so;

http {
  vhost_traffic_status_zone;

  server {
    listen 4444;
    server_name s1.com;
    location /metrics {
      limit_except GET {
        allow 10.99.0.0/16;
        deny  all;
      }
      vhost_traffic_status_display;
      vhost_traffic_status_display_format prometheus;
    }
  }

  server {
    vhost_traffic_status_filter_by_set_key $status $server_name;
    ...
  }
}

[u5surf]

@k0st1an Thanks. Unfortunately, we still cannot prepare the flag to disable the upstream statistics intentionally now.

We are fine what you said there are not good situations if we use the upstream directive we has enable its statistics simultaneously in some cases.

So I introduced the new parameter in #293 that can be through the upstream statistics. Can you review and try this parameter?

[k0st1an]

nginx version: nginx/1.24.0
built by gcc 12.2.0 (Debian 12.2.0-14)
built with OpenSSL 3.0.12 24 Oct 2023 (running with OpenSSL 3.0.11 19 Sep 2023)
TLS SNI support enabled
configure arguments: --prefix=/etc/nginx --sbin-path=/usr/sbin/nginx --modules-path=/usr/lib/nginx/modules --conf-path=/etc/nginx/nginx.conf --error-log-path=/var/log/nginx/error.log --http-log-path=/var/log/nginx/access.log --pid-path=/var/run/nginx.pid --lock-path=/var/run/nginx.lock --http-client-body-temp-path=/var/cache/nginx/client_temp --http-proxy-temp-path=/var/cache/nginx/proxy_temp --http-fastcgi-temp-path=/var/cache/nginx/fastcgi_temp --http-uwsgi-temp-path=/var/cache/nginx/uwsgi_temp --http-scgi-temp-path=/var/cache/nginx/scgi_temp --user=nginx --group=nginx --with-compat --with-file-aio --with-threads --with-http_addition_module --with-http_auth_request_module --with-http_dav_module --with-http_flv_module --with-http_gunzip_module --with-http_gzip_static_module --with-http_mp4_module --with-http_random_index_module --with-http_realip_module --with-http_secure_link_module --with-http_slice_module --with-http_ssl_module --with-http_stub_status_module --with-http_sub_module --with-http_v2_module --with-mail --with-mail_ssl_module --with-stream --with-stream_realip_module --with-stream_ssl_module --with-stream_ssl_preread_module --with-cc-opt='-g -O2 -ffile-prefix-map=/data/builder/debuild/nginx-1.24.0/debian/debuild-base/nginx-1.24.0=. -fstack-protector-strong -Wformat -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -fPIC' --with-ld-opt='-Wl,-z,relro -Wl,-z,now -Wl,--as-needed -pie' --add-module=/nginx-module-vts-issue-292

When two options are active, the statistics disappear

http {
  vhost_traffic_status_zone;
  vhost_traffic_status_bypass_upstream_stats on;  # <--

  server {
    listen 4444;
    server_name s1.com;
    location /metrics {
      limit_except GET {
        allow 10.99.0.0/16;
        deny  all;
      }
      vhost_traffic_status_display;
      vhost_traffic_status_display_format prometheus;
    }
  }

  server {
    vhost_traffic_status_filter_by_set_key $status $server_name;  # <--
    ...
  }
}

When only vhost_traffic_status_bypass_upstream_stats is set to on, there are no upstream statistics. But I don't see any difference in CPU load yet. I don't understand why this is so. Before that, I used https://deb.myguard.nl/nginx-modules/ repository. I will install it from this repository again and check my guess.

[k0st1an]

The problem reappeared with the https://deb.myguard.nl/nginx-modules/ repository. Configuration as in the first post. CPU throttle spike at 70k RPS

• 30 server (original nginx, without VTS) - 80k
• 31 server - 70k

$ dpkg -l | grep vhost
ii  libnginx-mod-http-vhost-traffic-status 1.25.3-4myguard21~bookworm              amd64        Nginx virtual host traffic status module

---

With upstreams disabled, the difference in CPU load is 2-3%. But I would like to test with the option vhost_traffic_status_status_filter_by_set_key $status $server_name;. But the stats are missing as I wrote above.

[u5surf]

@k0st1an It seems that I was misleaded your truely problem. So actually is your problem which you wanna really solve now the CPU utilization suddenly spiked? It is necessary for us to solve the problem if it can be reproduced absolutely the same situation for same configuration and the same request per sec. If it does that, we can get some profile with which we can investigate more such the situation.

https://gcc.gnu.org/onlinedocs/gcc-12.1.0/gcc/Instrumentation-Options.html