PHP 怎么PHP 性能指标

wen PHP项目 2026-07-29 1

本文目录导读：

PHP 怎么PHP 性能指标

核心性能指标
关键性能指标(KPI)
性能监控方案
性能优化指标
实战监控
最佳实践建议

我来系统性地介绍PHP的性能指标及其监控方法。

核心性能指标

执行时间

// 基础计时
$start = microtime(true);
// 你的代码
$end = microtime(true);
$executionTime = ($end - $start) * 1000; // 毫秒

内存使用

$memoryBefore = memory_get_usage();
// 你的代码
$memoryAfter = memory_get_usage();
$peakMemory = memory_get_peak_usage();

数据库查询性能

// 使用PDO示例
$db->setAttribute(PDO::ATTR_ERRMODE, PDO::ERRMODE_EXCEPTION);
$queryStart = microtime(true);
$result = $db->query("SELECT * FROM users");
$queryTime = microtime(true) - $queryStart;

关键性能指标(KPI)

响应时间指标

✅ 正常: < 200ms
⚠️ 警告: 200-500ms 
❌ 问题: > 1000ms

内存消耗指标

脚本执行: < 64MB
API请求: < 32MB
大型处理: < 256MB

吞吐量指标

QPS (每秒查询数)
- 小型站点: 100-500
- 中型站点: 500-2000
- 大型站点: 2000+

性能监控方案

内置监控类

class PerformanceMonitor {
    private $startTime;
    private $memoryStart;
    private $queries = [];
    public function __construct() {
        $this->startTime = microtime(true);
        $this->memoryStart = memory_get_usage();
    }
    public function getMetrics() {
        return [
            'execution_time' => (microtime(true) - $this->startTime) * 1000,
            'memory_usage' => memory_get_usage() - $this->memoryStart,
            'peak_memory' => memory_get_peak_usage(),
            'query_count' => count($this->queries),
            'total_query_time' => array_sum(array_column($this->queries, 'time'))
        ];
    }
}

Xdebug 分析

; php.ini 配置
xdebug.mode=profile
xdebug.output_dir=/tmp/profiler
xdebug.profiler_output_name=cachegrind.out.%p

Blackfire 工具

# 安装Blackfire
composer require blackfire/player
# 运行性能分析
blackfire run php script.php

性能优化指标

OpCache 状态

$opcacheStatus = opcache_get_status();
$metrics = [
    'hits' => $opcacheStatus['opcache_statistics']['hits'],
    'misses' => $opcacheStatus['opcache_statistics']['misses'],
    'cache_full' => $opcacheStatus['opcache_statistics']['oom_restarts'],
    'hit_rate' => ($hits / ($hits + $misses)) * 100
];

慢查询日志

-- MySQL 慢查询配置
SET GLOBAL slow_query_log = 'ON';
SET GLOBAL long_query_time = 1;
SET GLOBAL slow_query_log_file = '/var/log/mysql/slow.log';

PHP-FPM 指标

# 查看PHP-FPM状态
pm.status_path = /status
# 监控指标
pool:                 www
process manager:      dynamic
start time:          21/Oct/2024:10:00:00
start since:         3600
accepted conn:       15000
listen queue:        0
max listen queue:    5
listen queue len:    128
idle processes:      5
active processes:    10
total processes:     15
max active processes:20
max children reached:0
slow requests:       0

实战监控

请求级监控

class RequestMonitor {
    private $startTime;
    private $metrics = [];
    public function start() {
        $this->startTime = microtime(true);
        $this->metrics['memory_start'] = memory_get_usage();
        $this->metrics['queries'] = 0;
        $this->metrics['query_time'] = 0;
    }
    public function addQuery($time) {
        $this->metrics['queries']++;
        $this->metrics['query_time'] += $time;
    }
    public function finish() {
        $this->metrics['total_time'] = microtime(true) - $this->startTime;
        $this->metrics['memory_peak'] = memory_get_peak_usage();
        $this->metrics['memory_current'] = memory_get_usage();
        // 记录到日志
        $this->logMetrics();
    }
    private function logMetrics() {
        $log = sprintf(
            "Time: %.2fms | Mem: %.2fMB | Peak: %.2fMB | Queries: %d | QueryTime: %.2fms\n",
            $this->metrics['total_time'] * 1000,
            ($this->metrics['memory_current'] - $this->metrics['memory_start']) / 1048576,
            $this->metrics['memory_peak'] / 1048576,
            $this->metrics['queries'],
            $this->metrics['query_time'] * 1000
        );
        error_log($log, 3, '/var/log/php_monitor.log');
    }
}

性能报警配置

class PerformanceAlert {
    private $thresholds = [
        'execution_time' => 1000, // 毫秒
        'memory_usage' => 128 * 1024 * 1024, // 128MB
        'query_count' => 100,
        'query_time' => 5 // 秒
    ];
    public function check(array $metrics) {
        foreach ($this->thresholds as $metric => $threshold) {
            if (isset($metrics[$metric]) && $metrics[$metric] > $threshold) {
                $this->sendAlert($metric, $metrics[$metric], $threshold);
            }
        }
    }
    private function sendAlert($metric, $value, $threshold) {
        // 发送邮件、短信或写入告警日志
        error_log("ALERT: $metric exceeded threshold ($value > $threshold)");
    }
}

最佳实践建议

性能基准

页面加载: < 500ms
API响应: < 200ms
内存使用: < 64MB
数据库查询: < 50次/请求
文件包含: < 30个/请求

监控频率

线上环境: 实时监控
开发环境: 每次部署
压测环境: 每次变更

工具推荐

APM: New Relic, Datadog
Profiling: Xdebug, Blackfire, Tideways
日志: ELK Stack, Graylog
监控: Prometheus + Grafana

这些指标可以帮助你全面了解PHP应用的性能状况,及时发现问题并进行优化。