Laravel Octane Benchmark Tutorial

This guide helps you benchmark a Laravel application using three different runtime scenarios:

1. PHP-FPM (no Octane, no pooling)
2. Laravel Octane with Swoole (no persistent connections)
3. Laravel Octane with Swoole and MySQL persistent connections

---

🛠️ Prerequisites

• Laravel app installed
• MySQL running locally
• ApacheBench (ab) or wrk
• PHP >= 8.1
• Composer

---

🚀 Step 1: Install Laravel Octane

composer require laravel/octane
php artisan octane:install

Choose Swoole:

composer require open-swoole/swoole

Run Octane:

php artisan octane:start

---

🔧 Step 2: Enable Persistent Connection (Scenario C)

Edit config/database.php under the mysql connection:

'options' => extension_loaded('pdo_mysql') ? [
    PDO::ATTR_PERSISTENT => true,
] : [],

Clear config:

php artisan config:clear

---

🧪 Step 3: Create Benchmark Route

Edit routes/web.php:

use Illuminate\Support\Facades\DB;
use Illuminate\Support\Facades\Route;

Route::get('/benchmark', function () {
    $start = microtime(true);

    for ($i = 0; $i < 100; $i++) {
        DB::table('users')->limit(1)->get();
    }

    return [
        'duration_ms' => (microtime(true) - $start) * 1000,
        'memory' => memory_get_usage(true),
    ];
});

---

🧪 Step 4: Run Benchmarks

Use ApacheBench to run:

ab -n 1000 -c 100 http://127.0.0.1:8000/benchmark

Or use wrk:

wrk -t12 -c400 -d30s http://127.0.0.1:8000/benchmark

---

🔁 Step 5: Automate Benchmark

Use this bash script benchmark.sh to automate the tests:

chmod +x benchmark.sh
./benchmark.sh

---

📊 Sample Result Table

Scenario	Requests/sec	Avg Duration (ms)	Memory Usage	Notes
A (FPM)	~200	~500	High	Baseline
B (Octane)	~900	~80	Lower	Octane runtime
C (Octane + Pooling)	~1200	~50	Lowest	Fastest

---

✅ Tips

• Use php artisan octane:reload after code/config changes
• Avoid static/global state in Octane
• Preload services using Octane::warm()

---

📁 Files

• benchmark.sh: Benchmark automation script
• routes/web.php: Benchmark endpoint

---

🙋 Need Help?

Let me know if you want this exported as a GitHub repo or Gist!