Laravel Slower: Optimize Your DB Queries with AI

Laravel Slower is a powerful package designed for Laravel developers who want to enhance the performance of their applications. It intelligently identifies slow database queries and leverages AI to suggest optimal indexing strategies and other performance improvements. Whether you're debugging or routinely monitoring your application, Laravel Slower provides actionable insights to streamline database interactions.

Installation

You can install the package via composer:

composer require halilcosdu/laravel-slower

You can publish the config file with:

php artisan vendor:publish --tag="slower-config"

This is the contents of the published config file:

You can disable AI recommendations by setting the ai_recommendation key to false in the config file. If you disable AI recommendations, the package will not make any API requests to OpenAI.

<?php

// config for HalilCosdu/Slower

use HalilCosdu\Slower\Models\SlowLog;

return [
    'enabled' => env('SLOWER_ENABLED', true),
    'threshold' => env('SLOWER_THRESHOLD', 10000),
    'resources' => [
        'table_name' => (new SlowLog)->getTable(),
        'model' => SlowLog::class,
    ],
    'ai_recommendation' => env('SLOWER_AI_RECOMMENDATION', true),
    'recommendation_model' => env('SLOWER_AI_RECOMMENDATION_MODEL', 'gpt-4'),
    'recommendation_use_explain' => env('SLOWER_AI_RECOMMENDATION_USE_EXPLAIN', true),
    'ignore_explain_queries' => env('SLOWER_IGNORE_EXPLAIN_QUERIES', true),
    'ignore_insert_queries' => env('SLOWER_IGNORE_INSERT_QUERIES', true),
    'open_ai' => [
        'api_key' => env('OPENAI_API_KEY'),
        'organization' => env('OPENAI_ORGANIZATION'),
        'request_timeout' => env('OPENAI_TIMEOUT'),
    ],
    'prompt' => env('SLOWER_PROMPT', 'As a distinguished database optimization expert, your expertise is invaluable for refining SQL queries to achieve maximum efficiency. Schema json provide list of indexes and column definitions for each table in query. Also analyse the output of EXPLAIN ANALYSE and provide recommendations to optimize query. Please examine the SQL statement provided below including EXPLAIN ANALYSE query plan. Based on your analysis, could you recommend sophisticated indexing techniques or query modifications that could significantly improve performance and scalability?'),
];

You can publish and run the migrations with:

php artisan vendor:publish --tag="slower-migrations"

php artisan migrate

public function up()
{
    Schema::create(config('slower.resources.table_name'), function (Blueprint $table) {
        $table->id();
        $table->boolean('is_analyzed')->default(false)->index();
        $table->longtext('bindings');
        $table->longtext('sql');
        $table->float('time')->nullable()->index();
        $table->string('connection');
        $table->string('connection_name')->nullable();
        $table->longtext('raw_sql');
        $table->longtext('recommendation')->nullable();

        $table->timestamps();
    });
}

public function down(): void
{
    Schema::dropIfExists(config('slower.resources.table_name'));
}

Usage

You can register the commands with your scheduler.

php artisan slower:clean /*{days=15}  Delete records older than 15 days.*/
php artisan slower:analyze /*Analyze the records where is_analyzed=false*/

    use HalilCosdu\Slower\Commands\AnalyzeQuery;
    use HalilCosdu\Slower\Commands\SlowLogCleaner;

    protected $commands = [
        AnalyzeQuery::class,
        SlowLogCleaner::class,
    ];

    /**
     * Define the application's command schedule.
     */
    protected function schedule(Schedule $schedule): void
    {
        $schedule->command(AnalyzeQuery::class)->runInBackground()->daily();
        $schedule->command(SlowLogCleaner::class)->runInBackground()->daily();
    }

$model = \HalilCosdu\Slower\Models\SlowLog::first();

\HalilCosdu\Slower\Facades\Slower::analyze($model): Model;

dd($model->raw_sql); /*select count(*) as aggregate from "product_prices" where "product_id" = '1' and "price" = '0' and "discount_total" > '0'*/

dd($model->recommendation);

Example Screen

Example Recommendation

In order to improve database performance and scalability, here are some suggestions below:

1. Indexing: Effective database indexing can significantly speed up query performance. For your query, consider adding a combined (composite) index on product_id, price, and discount_total. This index would work well because the where clause covers all these columns.

CREATE INDEX idx_product_prices
ON product_prices (product_id, price, discount_total);

(Note: The order of the columns in the index might depend on the selectivity of the columns and the data distribution. Therefore, you might have to reorder them depending on your specific situation.)

2. Data Types: Ensure that the values being compared are of appropriate data types. Comparing or converting inappropriate data types at run time will slow down the search. It appears that you're using string comparisons ('1') for product_id, price, and discount_total which are likely numerical columns. Remove the quotes for these where clause conditions.

Updated Query:

SELECT COUNT(*) AS aggregate
FROM product_prices
WHERE product_id = 1
AND price = 0
AND discount_total > 0;

3. ANALYZE: Another practice to improve query performance could be running the ANALYZE command. This command collects statistics about the contents of tables in the database, and stores the results in the pg_statistic system catalog. Subsequently, the query planner uses these statistics to help determine the most efficient execution plans for queries.

ANALYZE product_prices;

Remember to periodically maintain your index to keep up with the CRUD operations that could lead to index fragmentation. Depending on your DBMS, you might want to REBUILD or REORGANIZE your indices.

Testing

composer test

Roadmap

• [ ] Create a documentation page.
• [ ] Begin development of version 2.
• [ ] Auto Indexer (Premium Feature)
• [ ] Create a FilamentPHP plugin.

Changelog

Please see CHANGELOG for more information on what has changed recently.

Contributing

Please see CONTRIBUTING for details.

Security Vulnerabilities

Please review our security policy on how to report security vulnerabilities.

Credits

• Halil Cosdu
• All Contributors

License

The MIT License (MIT). Please see License File for more information.