Boost Nginx Throughput 3×: 5 Essential Performance Tuning Parameters

Nginx is a core component of large‑scale architectures, and its performance often becomes the first bottleneck under high concurrency.

Five Key Tuning Parameters

worker_processes : Utilizes multiple CPU cores. Set to the number of cores, e.g., worker_processes auto; This raises CPU utilization from ~20% to 90% and increases throughput by about 1.5×.

worker_connections : Increases the maximum concurrent connections per worker. Example configuration:

events {
    worker_connections 65535;
}

The theoretical maximum connections equal worker_processes * worker_connections. Adjust ulimit -n accordingly to avoid "Too many open files" errors, yielding a ten‑fold increase in concurrent connections.

epoll : Enables Linux's high‑performance I/O multiplexing. Add to the events block:

events {
    use epoll;
}

Compared to select/poll (O(n)), epoll operates in O(1), cutting CPU usage by ~50% and reducing latency under heavy load.

sendfile & TCP optimizations : Zero‑copy file transmission and TCP tweaks improve I/O throughput.

sendfile on;
    tcp_nopush on;
    tcp_nodelay on;

These settings boost file transfer performance by ~30% and lower latency by ~15%.

keepalive optimization : Adjusts idle connection timeout and request limit.

keepalive_timeout 30s;
    keepalive_requests 1000;

Shortening idle time frees resources, while increasing request count per connection reduces handshake overhead.

Applying these configurations can increase overall Nginx throughput by more than 300% in typical high‑traffic environments.