Master Nginx: Comprehensive Guide to Reverse Proxy, Load Balancing, and Configuration

What can Nginx do?

1.1 Reverse proxy – Accept external requests, forward them to internal servers, and return the responses to clients.

1.2 Load balancing – Distribute requests among two or more backend servers using various strategies such as round‑robin (RR), weighted, ip_hash, fair (third‑party), and url_hash (third‑party).

1.3 HTTP server (static/dynamic separation) – Serve static resources directly and proxy dynamic API requests to backend services. Example configuration:

<span>server {</span>
<span>  listen 80;</span>
<span>  server_name localhost;</span>
<span>  client_max_body_size 1024M;</span>
<span>  location / {</span>
<span>    root e:/wwwroot;</span>
<span>    index index.html;</span>
<span>  }</span>
<span>}</span>

1.4 Forward proxy (less common) – Clients explicitly specify the target server; the proxy hides client details. Example configuration:

<span>server {</span>
<span>  listen 80;</span>
<span>  server_name localhost nginx.tangll.cn;</span>
<span>  resolver 8.8.8.8;</span>
<span>  location / {</span>
<span>    proxy_pass http://$http_host$request_uri;</span>
<span>  }</span>
<span>}</span>

Configuration Details

2.1 Overall structure

2.2 Global (main) block

<span># main – global configuration</span>
<span>user nobody nobody;</span>
<span>pid logs/nginx.pid;</span>
<span>worker_processes 2;</span>
<span>worker_rlimit_nofile 1024;</span>
<span>error_log logs/error.log info;</span>
<span>events { … }</span>
<span>http { … }</span>

Key directives:

user – set the user/group for worker processes.

pid – file storing the master process ID.

worker_processes – number of worker processes (usually a multiple of CPU cores).

worker_rlimit_nofile – maximum open files per process.

error_log – path and level of error logging.

2.3 Events block

<span>events {</span>
<span>  worker_connections 1024;</span>
<span>  multi_accept on;</span>
<span>  use epoll;</span>
<span>}</span>

Directives control connection handling and I/O model.

2.4 HTTP block

<span>http {</span>
<span>  sendfile on;</span>
<span>  tcp_nopush on;</span>
<span>  tcp_nodelay on;</span>
<span>  keepalive_timeout 65;</span>
<span>  send_timeout 50;</span>
<span>  client_max_body_size 10m;</span>
<span>  client_body_buffer_size 128k;</span>
<span>  types_hash_max_size 2048;</span>
<span>  include mime.types;</span>
<span>  default_type application/octet-stream;</span>
<span>  access_log logs/access.log main;</span>
<span>  error_log logs/error.log;</span>
<span>  gzip on;</span>
<span>  underscores_in_headers on;</span>
<span>  proxy_connect_timeout 500;</span>
<span>  proxy_read_timeout 600;</span>
<span>  proxy_send_timeout 500;</span>
<span>  proxy_buffer_size 128k;</span>
<span>  proxy_buffers 4 128k;</span>
<span>  proxy_busy_buffers_size 256k;</span>
<span>  proxy_temp_file_write_size 128k;</span>
<span>  proxy_temp_path d:/nginx-1.15.2/temp;</span>
<span>  proxy_cache_path d:/nginx-1.15.2/cache levels=1:2 keys_zone=cache_one:200m inactive=1d max_size=30g;</span>
<span>  upstream zuul_server { server 192.168.9.51:8000; }</span>
<span>  server {</span>
<span>    listen 8001;</span>
<span>    server_name localhost;</span>
<span>    location /report { alias D:/nginx-1.15.2/html/report-app/; }</span>
<span>    location / { alias D:/nginx-1.15.2/html/master-app/; }</span>
<span>    location /api/ { proxy_pass http://192.168.9.51:8057/; … }</span>
<span>  }</span>
<span>}</span>

2.4.4 Server block

listen – port to listen on (default 80).

server_name – hostnames for this virtual server.

root – document root for static files.

index – default index file.

access_log / error_log – log file locations.

2.4.5 Location module

Location directives match request URIs and can be precise (=), prefix (^~), or regex (~, ~*). Examples:

<span>location / { … }</span>
<span>location = / { … }</span>
<span>location ^~ /images/ { … }</span>
<span>location ~ \.(gif|jpg|jpeg)$ { … }</span>
<span>location ~* \.(css|js)$ { … }</span>

Directives inside location include root, alias, proxy_pass, and proxy_set_header.

2.5 Upstream module

Defines backend servers for load balancing. Example:

<span>upstream backend {</span>
<span>  server 192.168.1.10:8001;</span>
<span>  server 192.168.1.11:8001;</span>
<span>}</span>

Supports weight, max_fails, fail_timeout, backup, ip_hash, fair, url_hash, etc.

2.6 Stream block

<span>stream {</span>
<span>  upstream backend {</span>
<span>    hash $remote_addr consistent;</span>
<span>    server 192.168.9.51:8888 weight=5;</span>
<span>    server 192.168.8.230:8888 max_fails=3 fail_timeout=30s;</span>
<span>  }</span>
<span>  server {</span>
<span>    listen *:9999 so_keepalive=on;</span>
<span>    proxy_connect_timeout 1s;</span>
<span>    proxy_timeout 20s;</span>
<span>    proxy_pass backend;</span>
<span>    proxy_protocol on;</span>
<span>  }</span>
<span>}</span>

Provides TCP proxy capabilities.

For a complete reference, consult the official Nginx documentation.