Boost Nginx Performance: 10 Proven Configuration Tweaks

1. Event‑driven model

Nginx processes requests using an event‑driven architecture instead of spawning a process or thread per connection, allowing it to handle massive concurrency with minimal resource consumption.

2. Optimize worker processes

The worker_processes directive defines how many worker processes Nginx starts. Setting it to auto (or the number of CPU cores) fully utilizes the server’s CPU resources. The worker_connections directive controls the maximum concurrent connections per worker; a typical value is 1024 or higher.

worker_processes auto;
worker_connections 1024;

3. Enable gzip compression

Gzip reduces the amount of data sent to clients, speeding up page loads. Common settings include:

gzip on;
gzip_types text/plain text/css application/json application/javascript text/xml application/xml+rss;
gzip_comp_level 5;
gzip_buffers 16 8k;
gzip_vary on;
gzip_proxied any;

4. Static file caching

Configure appropriate HTTP cache headers so browsers cache static assets (images, CSS, JS, fonts) for a long period, reducing repeat requests.

location ~* \.(jpg|jpeg|png|gif|ico|css|js|woff|woff2|eot|ttf)$ {
    expires 30d;
    add_header Cache-Control "public, no-transform";
}

5. Cache dynamic content

Use fastcgi_cache or proxy_cache to store responses from back‑end applications (e.g., PHP‑FPM, Node.js). Define a cache zone, key, and validity period.

http {
    fastcgi_cache_path /var/cache/nginx levels=1:2 keys_zone=my_cache:10m inactive=60m;
    fastcgi_cache_key "$scheme$request_method$host$request_uri";
}
server {
    location ~ \.php$ {
        fastcgi_cache my_cache;
        fastcgi_cache_valid 200 60m;
        fastcgi_pass 127.0.0.1:9000;
    }
}

6. Reduce DNS lookup time

If upstream servers are referenced by domain name, Nginx resolves them at start‑up. Use the resolver directive with a reliable DNS server (e.g., 8.8.8.8) to avoid repeated lookups.

location / {
    resolver 8.8.8.8;
    proxy_pass http://backend_server;
}

7. Enable TCP_NODELAY

Setting tcp_nodelay on forces packets to be sent immediately instead of waiting to fill buffers, which is beneficial for low‑latency applications.

http {
    tcp_nodelay on;
}

8. Optimize logging

Logging incurs I/O overhead. Disable unnecessary access logs or buffer them in memory.

http {
    access_log off;               # disable access log
    # or
    # access_log /var/log/nginx/access.log buffer=32k;
}

9. Rate limiting

Use the limit_req_zone and limit_req modules to cap request rates per IP, protecting against abuse.

http {
    limit_req_zone $binary_remote_addr zone=one:10m rate=1r/s;
}
server {
    location /login/ {
        limit_req zone=one burst=5;
    }
}

10. Use HTTP/2

HTTP/2 provides multiplexing, server push, and header compression. Enable it by adding the http2 flag to the listen directive.

server {
    listen 443 ssl http2;
    # ... other settings ...
}