通常来说，一个优化良好的 Nginx Linux 服务器可以达到 500,000 – 600,000 次/秒 的请求处理性能，然而我的 Nginx 服务器可以稳定地达到 904,000 次/秒 的处理性能，并且我以此高负载测试超过 12 小时，服务器工作稳定。
# This number should be, at maximum, the number of CPU cores on your system.
# (since nginx doesn't benefit from more than one worker per CPU.)
# 这里的数值不能超过 CPU 的总核数，因为在单个核上部署超过 1 个 Nginx 服务进程并不起到提高性能的作用。
worker_processes 24;
# Number of file descriptors used for Nginx. This is set in the OS with 'ulimit -n 200000'
# or using /etc/security/limits.conf
# Nginx 最大可用文件描述符数量，同时需要配置操作系统的 "ulimit -n 200000"，或者在 /etc/security/limits.conf 中配置。
worker_rlimit_nofile 200000;
# only log critical errors
# 只记录 critical 级别的错误日志
error_log /var/log/nginx/error.log crit
# Determines how many clients will be served by each worker process.
# (Max clients = worker_connections * worker_processes)
# "Max clients" is also limited by the number of socket connections available on the system (~64k)
# 配置单个 Nginx 单个进程可服务的客户端数量，（最大值客户端数 = 单进程连接数 * 进程数 ）
# 最大客户端数同时也受操作系统 socket 连接数的影响（最大 64K ）
worker_connections 4000;
# essential for linux, optmized to serve many clients with each thread
# Linux 关键配置，允许单个线程处理多个客户端请求。
use epoll;
# Accept as many connections as possible, after nginx gets notification about a new connection.
# May flood worker_connections, if that option is set too low.
# 允许尽可能地处理更多的连接数，如果 worker_connections 配置太低，会产生大量的无效连接请求。
multi_accept on;
# Caches information about open FDs, freqently accessed files.
# Changing this setting, in my environment, brought performance up from 560k req/sec, to 904k req/sec.
# I recommend using some varient of these options, though not the specific values listed below.
# 缓存高频操作文件的FDs（文件描述符/文件句柄）
# 在我的设备环境中，通过修改以下配置，性能从 560k 请求/秒 提升到 904k 请求/秒。
# 我建议你对以下配置尝试不同的组合，而不是直接使用这几个数据。
open_file_cache max=200000 inactive=20s;
open_file_cache_valid 30s;
open_file_cache_min_uses 2;
open_file_cache_errors on;
# Buffer log writes to speed up IO, or disable them altogether
# 将日志写入高速 IO 存储设备，或者直接关闭日志。
# access_log /var/log/nginx/access.log main buffer=16k;
access_log off;
# Sendfile copies data between one FD and other from within the kernel.
# More efficient than read() + write(), since the requires transferring data to and from the user space.
# 开启 sendfile 选项，使用内核的 FD 文件传输功能，这个比在用户态用 read() + write() 的方式更加高效。
sendfile on;
# Tcp_nopush causes nginx to attempt to send its HTTP response head in one packet,
# instead of using partial frames. This is useful for prepending headers before calling sendfile,
# or for throughput optimization.
# 打开 tcp_nopush 选项，Nginux 允许将 HTTP 应答首部与数据内容在同一个报文中发出。
# 这个选项使服务器在 sendfile 时可以提前准备 HTTP 首部，能够达到优化吞吐的效果。
tcp_nopush on;
# don't buffer data-sends (disable Nagle algorithm). Good for sending frequent small bursts of data in real time.
# 不要缓存 data-sends （关闭 Nagle 算法），这个能够提高高频发送小数据报文的实时性。
tcp_nodelay on;
# Timeout for keep-alive connections. Server will close connections after this time.
# 配置连接 keep-alive 超时时间，服务器将在超时之后关闭相应的连接。
keepalive_timeout 30;
# Number of requests a client can make over the keep-alive connection. This is set high for testing.
# 单个客户端在 keep-alive 连接上可以发送的请求数量，在测试环境中，需要配置个比较大的值。
keepalive_requests 100000;
# allow the server to close the connection after a client stops responding. Frees up socket-associated memory.
# 允许服务器在客户端停止发送应答之后关闭连接，以便释放连接相应的 socket 内存开销。
reset_timedout_connection on;
# send the client a "request timed out" if the body is not loaded by this time. Default 60.
# 配置客户端数据请求超时时间，默认是 60 秒。
client_body_timeout 10;
# If the client stops reading data, free up the stale client connection after this much time. Default 60.
# 客户端数据读超时配置，客户端停止读取数据，超时时间后断开相应连接，默认是 60 秒。
send_timeout 2;
# Compression. Reduces the amount of data that needs to be transferred over the network
# 压缩参数配置，减少在网络上所传输的数据量。
gzip on;
gzip_min_length 10240;
gzip_proxied expired no-cache no-store private auth;
gzip_types text/plain text/css text/xml text/javascript application/x-javascript application/xml;
gzip_disable "MSIE [1-6].";
下面的配置项，我建议每次只修订其中一项，之后用网络性能测试工具 netperf、iperf 或是用我类似的测试脚本 cluster-netbench.pl 对服务器进行多次测试。
# Increase system IP port limits to allow for more connections
# 调高系统的 IP 以及端口数据限制，从可以接受更多的连接
net.ipv4.ip_local_port_range = 2000 65000
net.ipv4.tcp_window_scaling = 1
# number of packets to keep in backlog before the kernel starts dropping them
# 设置协议栈可以缓存的报文数阀值，超过阀值的报文将被内核丢弃
net.ipv4.tcp_max_syn_backlog = 3240000
# increase socket listen backlog
# 调高 socket 侦听数阀值
net.core.somaxconn = 3240000
net.ipv4.tcp_max_tw_buckets = 1440000
# Increase TCP buffer sizes
# 调大 TCP 存储大小
net.core.rmem_default = 8388608
net.core.rmem_max = 16777216
net.core.wmem_max = 16777216
net.ipv4.tcp_rmem = 4096 87380 16777216
net.ipv4.tcp_wmem = 4096 65536 16777216
net.ipv4.tcp_congestion_control = cubic
每次修订配置之后都需要执行以下命令使之生效.