当前位置：首页 > news >正文

Laraver SQL日志服务开发

news 来源：原创 2025/6/24 18:14:19

Laravel 项目运行中，有时候需要查看sql语句，分析sql运行的耗时，sql语句的复杂程度分析等等

总之，sql的执行在项目中，非常关键，接下来将说明在laravel 中配置一个sql日志记录服务。

开发思路：

使用 Laravel 事件系统监听 QueryExecuted 事件
通过监听服务，将监听到的sql信息记录到日志文件

代码开发

1. 日志通道配置一个记录sql的日志通道

文件：config/logging.php

'channels' => [
    'sqllog' => [
        'driver' => 'daily',
        'path' => storage_path('logs/sql.log'),
        'level' => 'debug',
        'days' => 7,
        'permission' => 0664,
    ],
];

通道名称 sqllog，驱动程序 daily , path是路径，level 为等级，days 保留7天，permission 为文件权限

2.事件系统配置监听事件

文件：app/Providers/EventServiceProvider.php

 protected $listen = [
        QueryExecuted::class => [
            LogDBQuery::class,
        ],
    ];

QueryExecuted::class 是 Laravel 数据库系统中的一个核心事件，也是 Laravel 数据库层最基础也是最有用的事件之一，其在数据库完成查询后触发。
- 触发的时机:
  - SELECT 查询
  - INSERT/UPDATE/DELETE 操作
  - 事务操作（BEGIN, COMMIT, ROLLBACK）
  - 存储过程调用
- 事件对象包含的信息
  - public string $sql; // 执行的SQL语句（带占位符）
  - public array $bindings; // 绑定的参数值
  - public float $time; // 执行时间（毫秒）
  - public string $connectionName; // 使用的连接名称（如’mysql’）
  - public object $connection; // 数据库连接实例
LogDBQuery::class 是自己开发一个监听服务类

3.监听服务开放

目录：app/Listeners，创建类 LogDBQuery.php

<?php

namespace App\Listeners;

use Illuminate\Contracts\Queue\ShouldQueue;
use Illuminate\Database\Events\QueryExecuted;
use Illuminate\Queue\InteractsWithQueue;
use Illuminate\Support\Facades\App;
use Illuminate\Support\Facades\Log;
use Illuminate\Support\Str;

class LogDBQuery
{
    // 慢查询阈值(毫秒)
    protected const SLOW_QUERY_THRESHOLD = 500;

    /**
     * 处理查询事件
     */
    public function handle(QueryExecuted $event): void
    {
        // 只在本地和测试环境记录
        if(!App::environment(['local', 'testing'])){
            return;
        }

        // 随机抽取 10% 的查询日志
        if(random_int(1, 100) > 10){
            return;
        }

        $context = $this->buildLogContext($event);

        if($event->time > self::SLOW_QUERY_THRESHOLD){
            Log::channel('sqllog')->warning("Slow query detected", $context);
        }else{
            Log::channel('sqllog')->debug("Query executed", $context);
        }
    }

    /**
     * 格式化 SQL 语句
     */
    protected function formatSql(string $sql, array $bindings): string
    {
        if(empty($bindings)){
            return $sql;
        }

        $escapedBindings = array_map(fn($value) => $this->escapeBinding($value), $bindings);

        return Str::replaceArray('?', $escapedBindings, $sql);
    }

    /**
     * 转义绑定参数
     */
    protected function escapeBinding(mixed $value): string
    {
        if(is_null($value)){
            return 'NULL';
        }

        if(is_bool($value)){
            return $value ? '1' : '0';
        }

        if(is_numeric($value)){
            return (string)$value;
        }

        return "'" . addslashes((string)$value) . "'";
    }

    /**
     * 构建日志上下文
     */
    protected function buildLogContext(QueryExecuted $event): array
    {
        return [
            'sql' => $this->formatSql($event->sql, $event->bindings),
            'time' => $event->time . 'ms',
            'connection' => $event->connectionName,
            'bindings' => $event->bindings,
            'raw_sql' => $event->sql,
        ];
    }
}

监听类执行流程

1事件触发 → 2. 框架查找监听器 → 3. 实例化监听器 → 4. 调用 handle()

代码解释：

SLOW_QUERY_THRESHOLD 添加一个慢查询的阈值，如果查询时间大于阈值，则是慢查询
handle 自动执行：

当非生产环境，则执行

 if(!App::environment(['local', 'testing'])){
    return;
  }

随机抽查 10%的日志记录

 if(random_int(1, 100) > 10){
    return;
   }

构建日志上下文数组，记录sql的语句，时间，连接等信息

protected function buildLogContext(QueryExecuted $event): array
    {
        return [
            'sql' => $this->formatSql($event->sql, $event->bindings),
            'time' => $event->time . 'ms',
            'connection' => $event->connectionName,
            'bindings' => $event->bindings,
            'raw_sql' => $event->sql,
        ];
    }

检测是否是慢查询，记录不同级别的日志，提高日志的可观察性

if($event->time > self::SLOW_QUERY_THRESHOLD){
     Log::channel('sqllog')->warning("慢查询SQL", $context);
 }else{
     Log::channel('sqllog')->debug("执行SQL", $context);
 }

完成上述构建，测试一下：

[2025-03-31 14:10:26] local.DEBUG: 执行SQL {"sql":"select count(*) as aggregate from `project_info` where `project_info`.`deleted_at` is null","time":"24.21ms","connection":"mysql","bindings":[],"raw_sql":"select count(*) as aggregate from `project_info` where `project_info`.`deleted_at` is null"} 
[2025-03-31 14:10:26] local.DEBUG: 执行SQL {"sql":"select * from `project_info` where `project_info`.`deleted_at` is null order by `id` desc limit 10 offset 0","time":"24.77ms","connection":"mysql","bindings":[],"raw_sql":"select * from `project_info` where `project_info`.`deleted_at` is null order by `id` desc limit 10 offset 0"}