README

Measure peak memory and execution time in your Laravel app — for HTTP requests and Artisan commands. Ships with headers, logs, and optional Telescope integration.

Features

• 🔎 Precise metrics: duration (ms), start/end/peak memory and memory diff.
• 🌐 HTTP profiling via middleware + optional response headers: X-Memory-Peak, X-Duration-ms.
• 🧰 Artisan profiling via a simple trait wrapper.
• 🧱 Manual spans (start/stop) via the service or facade.
• 🪵 Pluggable outputs: app log, Laravel Telescope (or both).
• 🎚️ Threshold: only log when peak memory exceeds N bytes.

Requirements

• PHP >= 8.1
• Laravel 10.x / 11.x / 12.x

Installation

composer require phpner/laravel-profiler
php artisan vendor:publish --tag=config --provider="Phpner\\MemoryProfiler\\MemoryProfilerServiceProvider"

This publishes config/laravel-memory-profiler.php.

Quick start

HTTP requests

Register the middleware (global or route group):

// app/Http/Kernel.php
protected $middleware = [
    // ...
    \Phpner\MemoryProfiler\Middleware\HttpProfiler::class,
];

When enabled, each request will:

• write a log entry with metrics and request context;
• (optionally) add headers X-Memory-Peak and X-Duration-ms to the response.

Artisan commands

Wrap your command logic with the provided trait:

use Illuminate\Console\Command;
use Phpner\MemoryProfiler\Console\Concerns\ProfilesMemory;

class ImportUsers extends Command
{
    use ProfilesMemory;

    protected $signature = 'users:import';

    public function handle(): int
    {
        return $this->withMemoryProfiling(function () {
            // your command logic
            $this->info('Importing...');
            // ...
            return 0; // exit code
        });
    }
}

The profiler will log command name and exit code along with metrics.

Manual spans

Use the service (DI) or the facade to measure any piece of code.

use Phpner\MemoryProfiler\MemoryProfiler; // service
use Phpner\MemoryProfiler\Facades\MemoryProfiler as MP; // facade

// 1) Service
public function show(MemoryProfiler $profiler)
{
    $ctx = $profiler->start('code');
    // ... your code ...
    $profiler->stop($ctx, [
        'label' => 'users.show',
        'meta'  => ['id' => 42],
    ]);
}

// 2) Facade
$ctx = MP::start('code');
// expensive work
MP::stop($ctx, ['label' => 'expensive.segment']);

Configuration

config/laravel-memory-profiler.php:

return [
    'enabled'          => env('MEMORY_PROFILER_ENABLED', true),
    'driver'           => env('MEMORY_PROFILER_DRIVER', 'log'), // log | telescope | both
    'threshold'        => env('MEMORY_PROFILER_THRESHOLD', 0),  // bytes; 0 = log everything
    'response_headers' => env('MEMORY_PROFILER_HEADERS', true),
];

Common .env options

MEMORY_PROFILER_ENABLED=true
MEMORY_PROFILER_DRIVER=both     # log | telescope | both
MEMORY_PROFILER_THRESHOLD=0     # e.g. 2097152 (2 MB)
MEMORY_PROFILER_HEADERS=true

Output

Response headers (HTTP)

X-Memory-Peak: 1048576
X-Duration-ms: 123

Log entry (example)

{
  "duration_ms": 123,
  "memory_start": 524288,
  "memory_end": 786432,
  "memory_peak": 1048576,
  "memory_diff": 262144,
  "type": "http",          // or "artisan" / "code"
  "label": "users.show",   // when provided for manual spans
  "method": "GET",
  "path": "api/users/42",
  "route": "users.show",
  "status": 200
}

Telescope (optional)

Set MEMORY_PROFILER_DRIVER=telescope or both. Entries appear in Telescope Log section with the memory tag.

Why this package?

• Lightweight and framework-native.
• Zero vendor lock‑in: works with logs you already collect; Telescope is optional.
• Transparent: you decide where to profile (global, group, specific commands, or manual spans).

Tips

• Start with threshold=0 in dev to see everything; raise it in prod to reduce noise.
• Put the middleware only on routes you care about if you want a narrower focus.
• Pair with request IDs/correlation IDs in your logger to join app logs and memory profiles.

Roadmap

• Queue job middleware.
• Blade/Laravel Debugbar-like panel.
• Per-route configuration overrides.

Contributing

PRs and issues are welcome! Please include tests where it makes sense.

License

Released under the MIT License.