Nginx 响应慢

[uwsgi] pythOnpath=/xxx static-map=/static=/xxx/static chdir=/xxx env=DJANGO_SETTINGS_MODULE=conf.settings module=xxx.wsgi master=True pidfile=logs/xxx.pid vacuum=True max-requests=100000 enable-threads=true processes=16 threads=32 listen=1024 log-slow=3000 daemOnize=logs/wsgi.log stats=/tmp/xxx/socket/stats.socket http=0.0.0.0:6187 buffer-size=220000000 socket-timeout=1500 harakiri=1500 http-timeout=1500

reqeust log

[pid: 10550|app: 0|req: 549/6061] 103.218.240.105 () {50 vars in 1037 bytes} [Mon Mar 8 15:24:30 2021] GET /api/v2/analysis/xxxx => generated 3890508 bytes in 397 msecs (HTTP/1.1 200) 5 headers in 222 bytes (1 switches on core 16)

2

nginx.conf

worker_processes 12; events { use epoll; worker_connections 65535; } http { include mime.types; include log_format.conf; include upstream.conf; default_type application/octet-stream; sendfile on; tcp_nopush on; keepalive_timeout 1800; server_tokens off; client_max_body_size 100m; gzip on; gzip_min_length 1k; gzip_buffers 4 16k; gzip_comp_level 5; gzip_types text/plain application/json application/Javascript application/x-Javascript text/css application/xml text/Javascript application/x-httpd-php image/jpeg image/gif image/png; gzip_vary off; include "site-enabled/*.conf"; }

upstream.conf

 upstream bv_crm_server_proxy_line { server proxy.xxxx.cn:6187 weight=100 fail_timeout=0; keepalive 500; }

log_format.conf

log_format upstream '$remote_addr - $host [$time_local] "$request" ' '$status $body_bytes_sent $request_time $upstream_response_time ' '"$http_user_agent" "$http_x_forwarded_for" ';

site-enabled.xxx.conf

server { listen 7020; server_name xxxx.xx.cn; client_max_body_size 100M; access_log logs/xxx.log upstream; root /home/smb/web/xxx/dist; client_header_buffer_size 16k; large_client_header_buffers 4 16k; location ^~ /api/ { proxy_redirect off; proxy_set_header Host $host; proxy_set_header X-Real-IP $remote_addr; proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for; proxy_send_timeout 1800; proxy_connect_timeout 1800; proxy_read_timeout 1800; proxy_ignore_client_abort on; proxy_pass http://bv_crm_server_proxy_line; } location / { try_files $uri /index.html =404; } }

192.168.12.12 - xxx.cn [08/Mar/2021:15:24:34 +0800] "GET /api/v2/analysis/xxx HTTP/1.1" 200 531500 4.714 4.714 "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.190 Safari/537.36" "103.120.18.243"

现在 nginx 的响应速度很慢，请求大佬帮忙看一下是不是哪里配置的有问题。

感谢

8 条回复 2021-03-09 11:31:27 +08:00

brader

2021-03-09 10:51:55 +08:00

你可否先尝试排除 nginx 之外的问题，仅仅测试 nginx ？
```
location /{
default_type text/plain;
return 200 "hello nginx!\n";
}
```

defunct9

2021-03-09 11:00:23 +08:00

upstream fastcgi_backend {
server 127.0.0.1:9000;

keepalive 8;
}

server {
...

location /fastcgi/ {
fastcgi_pass fastcgi_backend;
fastcgi_keep_conn on;
...
}
}

vegetableChick

2021-03-09 11:06:45 +08:00

@brader 线上的机器感觉没法试。。其他请求 nginx 响应是正常的, 这个请求返回的是一个 3m 左右的 json，不知道和这个有没有关系

vegetableChick

2021-03-09 11:09:35 +08:00

@defunct9 感谢回复，我没看懂您这个配置的意思。。可以大概说一下不

chendy

2021-03-09 11:16:14 +08:00

服务器带宽多大？ 3m 的 json，10m 的机器要差不多 3s 才能发完

barrysn

2021-03-09 11:23:14 +08:00

设置日志格式，看看 nginx 日志里记录的时间哪个长，
先确认问题出在哪里

$request_time Full request time, starting when NGINX reads the first byte from the client and ending when NGINX sends the last byte of the response body
$upstream_connect_time Time spent establishing a connection with an upstream server
$upstream_header_time Time between establishing a connection to an upstream server and receiving the first byte of the response header
$upstream_response_time Time between establishing a connection to an upstream server and receiving the last byte of the response body

brader

2021-03-09 11:30:32 +08:00

@vegetableChick 可以测试的，你可以不干涉你原有的东西，另增加一个匹配规则来测试就好了。
如果你返回一个 3m 大小的东西，有一定访问人数，访问频繁，服务器带宽不够的话，确实会非常慢的，打个比方，你这就好像是，分发下载 app，不使用 oss 来分发，而使用服务器带宽硬抗。

defunct9

2021-03-09 11:31:27 +08:00

o , 你这用的是 uwsgi，我建议换 fastcgi，用 fastcgi 的 keepalive 特性。加快速度。
而且呢
#include uwsgi_params;
#uwsgi_pass unix:///var/www/script/uwsgi.sock; # 指定 uwsgi 的 sock 文件所有动态请求
用这种会更快。
anyway，开 ssh，让我上去看看