Awesome

Laravel Indexer

Laravel Indexer monitors SELECT queries running on a page and allows to add database indexes to SELECT queries on the fly. It then presents results of EXPLAIN or MySQL's execution plan right on the page. The results presented by Indexer will help you see which indexes work best for different queries running on a page.

Indexes added by Indexer are automatically removed after results are collected while keeping your existing indexes intact.

CAUTION: PLEASE DO NOT USE THIS PACKAGE ON PRODUCTION! Since this package adds indexes to database on the fly, it is strongly recommended NOT to use this package in your production environment.

Note Since indexes are added and then removed dynamically to generate results, pages will load slow.

Requirements

• PHP >= 7
• Laravel 5.3+ | 6

Installation

Install via composer

composer require sarfraznawaz2005/indexer --dev

For Laravel < 5.5:

Add Service Provider to config/app.php in providers section

Sarfraznawaz2005\Indexer\ServiceProvider::class,

Publish package's config file by running below command:

php artisan vendor:publish --provider="Sarfraznawaz2005\Indexer\ServiceProvider"

It should publish config/indexer.php config file.

Screenshot

When enabled, you will see yellow/green/red box on bottom right:

• Yellow by default or when queries need to be optimized
• Green when total queries count matches optimized queries count.
• Red when one or more slow queries found and need to be optimized.

Config

enabled : Enable or disable Indexer. By default it is disabled.

check_ajax_requests : Specify whether to check queries in ajax requests.

ignore_tables : When you don't use watched_tables option, Indexer watches all tables. Using this option, you can ignore specified tables to be watched.

ignore_paths : These paths/patterns will NOT be handled by Indexer.

slow_time : Time in ms when queries will be considered slow.

output_to : Outputs results to given classes. By default Web class is included.

watched_tables : DB tables to be watched by Indexer. Here is example:

'watched_tables' => [
    'users' => [
        // list of already existing indexes to try
        'try_table_indexes' => ['email'],
        // new indexes to try
        'try_indexes' => ['name'],
        // new composite indexes to try
        'try_composite_indexes' => [
            ['name', 'email'],
        ],
    ],
],

• Here queries involving users DB table will be watched by Indexer.
  • try_table_indexes contains index names that you have already applied to your DB table. Indexer will simply try out your existing indexes to show EXPLAIN results. In this case, email index already exists in users table.
  • try_indexes can be used to add new indexes on the fly to DB table. In this case, name index will be added on the fly by Indexer and results will be shown of how that index performed.
  • Like try_indexes the try_composite_indexes can also be used to add composite indexes on the fly to DB table. In this case, composite index consisting of name and email will be added on the fly by Indexer and results will be shown of how that index performed.

Modes

Indexer can be used in following ways:

All Indexes Added By Indexer

Don't put any indexes manually on your tables instead let Indexer add indexes on the fly via try_indexes and/or try_composite_indexes options. Indexes added via these two options are automatically removed.

In this mode, you can actually see which indexes work best without actually applying on your tables. You can skip using try_table_indexes option in this case.

Already Present Indexes + Indexes Added By Indexer

You might have some indexes already present on your tables but you want to try out more indexes on the fly without actually adding those to the table. To specify table's existing indexes, use try_table_indexes option as mentioned earlier. And to try out new indexes on the fly, use try_indexes and/or try_composite_indexes options. Table's existing indexes (specified in try_table_indexes) will remain intact but indexes added via try_indexes and try_composite_indexes will be automatically removed.

Already Present Indexes

When you don't want Indexer to add any indexes on the fly and you have already specified indexes on your tables and you just want to see EXPLAIN results for specific tables for your indexes, in this case simply use try_table_indexes option only. Example:

'watched_tables' => [
    'users' => [
        'try_table_indexes' => ['email'],
    ],
    'posts' => [
        'try_table_indexes' => ['title'],
    ]
],

In this case, both email and title indexes are supposed to be already added to table manually.

No Indexes, Just Show EXPLAIN results for all SELECT queries

While previous three modes allow you to work with specific tables and indexes, you can use this mode to just show EXPLAIN results for all SELECT queries running on a page without adding any indexes on the fly. To use this mode, simply don't specify any tables in watched_tables option. If you don't want to include some tables in this mode, use ignore_tables option.

Misc

• Color of Indexer box on bottom right or query sections inside results changes to green if it finds query's EXPLAIN result has key present eg query actually used a key. This can be changed by creating your own function in your codebase called indexerOptimizedKeyCustom(array $queries) instead of default one indexerOptimizedKey which is present in file src/Helpers.php. Similarly, for ajax requests, you should define your own function called indexerOptimizedKeyCustom(explain_result). Here is example of each:

// php
function indexerOptimizedKeyCustom(array $query): string
{
   return trim($query['explain_result']['key']);
}

// javascript
function indexerOptimizedKeyCustom(explain_result) {
    return explain_result['key'] && explain_result['key'].trim();
}

Note: If Indexer has found any slow queries (enabled via slow_time option), the color of box on bottom right will always be red until you fix slow queries.

Limitation

• Indexer tries to find out tables names after FROM keyword in queries, therefore it cannot work with complex queries or ones that don't have table name after FROM keyword.

Security

If you discover any security related issues, please email sarfraznawaz2005@gmail.com instead of using the issue tracker.

Credits

• Sarfraz Ahmed
• All Contributors

License

Please see the license file for more information.