Caching Basics

Caching patterns: page/data caching, TTL, stampede control, and invalidation basics.

On this page

What Caching Solves

Caching reduces repeated work: DB queries, rendering, expensive computations. Done right, it improves latency and reduces load. Done wrong, it serves stale data or creates hard-to-debug behavior.

What to Cache

Good candidates: read-heavy data (categories), computed results, rendered fragments, and expensive API calls. Avoid caching user-specific sensitive responses unless you understand the scope.

Simple In-memory Cache (Educational)

This shows the concept. In production you might use Redis or APCu depending on environment.

<?php
$cache = [];

function cacheGet(array $cache, string $key): ?string {
  return $cache[$key] ?? null;
}

function cacheSet(array &$cache, string $key, string $value): void {
  $cache[$key] = $value;
}

TTL and Cache Keys

Use stable keys and explicit TTL. Keys must include parameters that change the result (page, language, filters).

<?php
$key = "items:page=" . (int)($_GET["page"] ?? 1) . ":lang=php";
$ttlSeconds = 60;

Stampede Control (Baseline)

Cache stampede happens when many requests miss at the same time and all recompute. A basic mitigation is short TTLs + jitter and optionally a lock in shared cache.

<?php
$ttl = 60;
$jitter = random_int(0, 10);
$effectiveTtl = $ttl + $jitter;

Production Tip

Plan invalidation: time-based TTL is simplest, event-based invalidation is harder but more correct. Start with small TTL + observability, then evolve to Redis cache with clear key strategy.

← Logging

OPcache →