Laminas Cache Storage Adapter Benchmark Laravel Package

Getting Started

1. Install: Run composer require laminas/laminas-cache-storage-adapter-benchmark — it’s a dev/test dependency, so consider --dev.
2. First use case: Wrap your existing cache adapter (e.g., Filesystem, Redis) with the Benchmark adapter to measure real overhead without changing your app logic.

   $benchmarkAdapter = new \Laminas\Cache\Storage\Adapter\Benchmark(
       new \Laminas\Cache\Storage\Adapter\Filesystem(['CacheDir' => '/tmp/cache'])
   );
   $storage = \Laminas\Cache\StorageFactory::factory($benchmarkAdapter);
   

3. Check metrics: After using $storage->setItem('key', 'val'), retrieve stats with $benchmarkAdapter->getStats() — look for write_time, read_time, count_write, count_read.

Implementation Patterns

• Configuration-driven benchmarking: In config/autoload/cache.local.php, dynamically inject the benchmark adapter in dev/test only:

  'caches' => [
      'default' => [
          'adapter' => [
              'name' => 'Benchmark',
              'options' => [
                  'adapter' => [
                      'name' => 'Filesystem',
                      'options' => ['CacheDir' => '/tmp/cache'],
                  ],
              ],
          ],
      ],
  ],
  

• Profiling middleware: Use the benchmark adapter inside a PSR-15 or Expressive-style middleware to log per-request cache metrics (e.g., in a debug toolbar or /debug/cache route).
• Side-by-side comparisons: Swap underlying adapters (e.g., Filesystem vs Apcu vs Redis) while keeping the benchmark wrapper — compare total_read_time, hit_ratio, and operation_count across runs.
• Automated baselining: Run CLI benchmarks (e.g., bin/benchmark-cache) to capture getStats() before/after refactoring or config tweaks, committing results to a baseline file.

Gotchas and Tips

• ⚠️ Never use in production: The adapter wraps every call in microtime(true) — significant overhead. Use environment guards:

  if (getenv('APP_ENV') === 'dev') {
      $adapter = new Benchmark($realAdapter);
  } else {
      $adapter = $realAdapter;
  }
  

• Nested adapters: Avoid double-wrapping (new Benchmark(new Benchmark($adapter))) — stats will break. Ensure only one benchmark adapter is active.
• Metrics granularity: getStats() returns cumulative values — reset between test runs using resetStats() (if available) or re-instantiate.
• Integration gotcha: The benchmark adapter does not implement all storage interface methods explicitly — rely on __call() forwarding. Confirm edge-case calls (e.g., normalizeOptions, hasOptions) still work if customized.
• Debug tip: Enable verbosity in your inner adapter (e.g., Filesystem::setTtl(0)) to isolate benchmark noise from backend latency.