I got a call from a new client last month. They ran a pretty successful WooCommerce shop, but the site crashed every single time they put on a sale. Their previous developer had installed a caching plugin and called it a day, and the server still timed out under any real load. The client was frustrated and losing money.
The problem wasn’t the plugin. It was that their whole WordPress server stack sat on a 10-year-old foundation. I’ve seen this a dozen times. You can’t bolt a turbo onto a lawnmower engine and expect it to win a race.
Your stack is more than a caching plugin
Years ago, a basic LAMP stack (Linux, Apache, MySQL, PHP) was all you needed. But traffic and user expectations have changed. Apache is fine, but it struggles with a lot of simultaneous connections, which is exactly what happens during a flash sale. Nginx handles that concurrency much better, so for high-traffic sites I default to it.
Now, about that caching plugin. My first move, a little embarrassing in hindsight, was to check its configuration. I spent a solid hour tweaking settings, figuring the last guy had set it up wrong. I got a small performance boost, right up until my load test brought the site down again. That told me the issue was foundational: you can’t fix a server problem with a PHP solution.
There is a second half to this, though: page caching only gets you so far. Every time a user logs in, adds something to their cart, or does anything dynamic, WordPress has to hit the database, and those queries are expensive. A persistent object cache helps here. It keeps the results of common queries in memory so PHP doesn’t have to keep hitting the database. The usual choice is Redis.
// Add this to your wp-config.php to integrate Redis
define( 'WP_REDIS_HOST', '127.0.0.1' );
define( 'WP_REDIS_PORT', 6379 );
define( 'WP_CACHE_KEY_SALT', 'your_unique_prefix' );
// Then you need a plugin like "Redis Object Cache" to connect it.How we actually fixed it
There isn’t one fix. You build a stack where each layer supports the others. For that client, we moved to a server running Nginx and the latest stable PHP version, and set up Redis for object caching. The site now holds up fine through traffic spikes. As Carl Alexander noted in his breakdown of the modern server stack, the pieces have to fit together for the site to be fast.
- Web Server: Use Nginx, not Apache.
- Page Cache: Handle it at the server level (Nginx FastCGI cache) before the request even hits PHP.
- Object Cache: Use a persistent cache like Redis to reduce database load.
Look, this stuff gets complicated fast. If you’re tired of debugging someone else’s mess and just want your site to work, drop my team a line. We’ve probably seen it before.
What does your server stack look like? Are you still running on a basic LAMP setup or have you made the switch?
