Unlock 10× Faster Responses: Inside Nginx’s Caching Mechanism

Nginx is a core component of large‑scale architectures, and its caching feature is one of the most important optimizations for high‑concurrency, high‑performance systems.

When configured properly, Nginx cache can dramatically reduce backend server load and lower response times, and in certain read‑heavy static or semi‑static scenarios it can achieve performance improvements of more than ten times.

There are two layers of caching related to Nginx:

Browser cache, which relies on Cache-Control, Expires, ETag, and Last-Modified mechanisms. A hit may return the resource locally or perform a 304 conditional validation.

Nginx proxy cache, which stores the complete upstream response on local disk or in memory and serves subsequent identical requests directly from Nginx without contacting the backend application.

The proxy cache operates in reverse‑proxy mode. The overall flow is:

Client request arrives at Nginx.

Nginx checks the cache according to the configured policy.

If the request hits the cache, Nginx returns the cached content immediately.

If the request misses, Nginx forwards it to the backend server, writes the upstream response into the cache, and then returns the response to the client.

To keep the cache effective, Nginx supports cache expiration times, update strategies, and cleaning mechanisms, allowing a balance between data consistency and performance gains.

Nginx cache configuration

The key to Nginx cache configuration lies in the design of parameters.

Important directives include: proxy_cache_path – defines the cache storage path, hierarchy levels, and memory index area. proxy_cache – enables caching within a specific server or location block. proxy_cache_valid – sets cache duration for different response status codes.

Example configuration:

server {
    listen 80;
    server_name example.com;

    location / {
        proxy_pass http://backend;   # upstream server
        proxy_cache my_cache;        # enable cache, specify zone
        proxy_cache_valid 200 302 10m;   # cache successful responses for 10 minutes
        proxy_cache_valid 404 1m;       # cache 404 responses for 1 minute
        proxy_cache_valid any 5m;        # default cache time for other statuses
        proxy_cache_use_stale error timeout updating http_500 http_502 http_503 http_504;
        proxy_cache_lock on;            # prevent cache stampede
        proxy_cache_bypass $http_cache_purge;  # allow forced refresh
        add_header -Cache-Status $upstream_cache_status;  # show HIT/MISS/BYPASS
    }
}

The above directives constitute a practical cache setup that maximizes hit rates while protecting against cache stampede and providing visibility into cache status.