// Micrometer 中创建 Counter
@Component
public class OrderMetrics {

    private final Counter orderCreatedCounter;
    private final Counter orderFailedCounter;

    public OrderMetrics(MeterRegistry registry) {
        this.orderCreatedCounter = Counter.builder("orders.created.total")
                .description("订单创建总数")
                .tag("service", "order-service")
                .register(registry);

        this.orderFailedCounter = Counter.builder("orders.failed.total")
                .description("订单创建失败总数")
                .tag("service", "order-service")
                .register(registry);
    }

    public void incrementOrderCreated() {
        orderCreatedCounter.increment();  // +1
    }

    public void incrementOrderFailed(String reason) {
        orderFailedCounter.increment();   // +1——带 tag 区分原因
    }
}

@Component
public class OrderMetrics {

    private final AtomicInteger pendingOrders = new AtomicInteger(0);

    public OrderMetrics(MeterRegistry registry) {
        Gauge.builder("orders.pending", pendingOrders, AtomicInteger::get)
                .description("当前待处理的订单数")
                .tag("service", "order-service")
                .register(registry);
    }

    public void incrementPending() { pendingOrders.incrementAndGet(); }
    public void decrementPending() { pendingOrders.decrementAndGet(); }
}

// 方式一：用 @Timed 注解——最简单
@RestController
@RequestMapping("/api/orders")
public class OrderController {

    @PostMapping
    @Timed(value = "orders.create.duration", description = "创建订单耗时")
    public Order createOrder(@RequestBody CreateOrderRequest request) {
        return orderService.createOrder(request);
    }
}

// 方式二：用 Timer——手动记录
@Component
public class OrderMetrics {

    private final Timer orderCreateTimer;

    public OrderMetrics(MeterRegistry registry) {
        this.orderCreateTimer = Timer.builder("orders.create.duration")
                .description("创建订单耗时")
                .publishPercentiles(0.5, 0.95, 0.99)  // 自动算 P50/P95/P99
                .sla(Duration.ofMillis(100), Duration.ofMillis(500))  // SLA 桶
                .register(registry);
    }

    public void recordOrderCreate(long durationMs) {
        orderCreateTimer.record(durationMs, TimeUnit.MILLISECONDS);
    }
}

version: '3.8'
services:

  prometheus:
    image: prom/prometheus:v2.48.0
    container_name: prometheus
    ports:
      - "9090:9090"
    volumes:
      - ./prometheus.yml:/etc/prometheus/prometheus.yml
      - prometheus-data:/prometheus
    command:
      - '--config.file=/etc/prometheus/prometheus.yml'
      - '--storage.tsdb.path=/prometheus'
      - '--storage.tsdb.retention.time=15d'    # 数据保留 15 天
      - '--web.enable-lifecycle'               # 允许热加载配置

  grafana:
    image: grafana/grafana:10.2.0
    container_name: grafana
    ports:
      - "3000:3000"
    environment:
      - GF_SECURITY_ADMIN_PASSWORD=admin123
      - GF_INSTALL_PLUGINS=grafana-clock-panel
    volumes:
      - grafana-data:/var/lib/grafana
    depends_on:
      - prometheus

volumes:
  prometheus-data:
  grafana-data:

# prometheus.yml
global:
  scrape_interval: 15s          # 每 15s 抓一次——生产默认
  evaluation_interval: 15s      # 每 15s 评估一次告警规则

# 抓取目标配置
scrape_configs:
  # Prometheus 自身的指标
  - job_name: 'prometheus'
    static_configs:
      - targets: ['localhost:9090']

  # Spring Boot 微服务——通过 Actuator 暴露
  - job_name: 'spring-boot-apps'
    metrics_path: '/actuator/prometheus'
    static_configs:
      - targets:
          - 'user-service:8081'
          - 'order-service:8082'
          - 'product-service:8083'
        labels:
          env: 'production'

<dependency>
    <groupId>org.springframework.boot</groupId>
    <artifactId>spring-boot-starter-actuator</artifactId>
</dependency>
<dependency>
    <groupId>io.micrometer</groupId>
    <artifactId>micrometer-registry-prometheus</artifactId>
</dependency>

# application.yml
management:
  endpoints:
    web:
      exposure:
        include: health,info,metrics,prometheus  # ← 暴露 /actuator/prometheus
  metrics:
    tags:
      application: ${spring.application.name}     # 给所有指标打上应用 Tag
    export:
      prometheus:
        enabled: true

# 验证——访问这个 URL 看到 Prometheus 格式的指标
curl http://localhost:8081/actuator/prometheus

# 输出示例：
# HELP jvm_memory_used_bytes The amount of used memory
# TYPE jvm_memory_used_bytes gauge
# jvm_memory_used_bytes{area="heap",application="user-service"} 1.5E8
# HELP http_server_requests_seconds Histogram of HTTP request durations
# TYPE http_server_requests_seconds histogram
# http_server_requests_seconds_bucket{method="GET",outcome="SUCCESS",status="200",uri="/api/users/{id}",le="0.1",} 120.0

# ① 直接查值——加 {label="value"} 过滤
http_server_requests_seconds_count{application="user-service", uri="/api/users/{id}"}

# ② rate()——计算每秒增长速率（Counter 用）
# Counter 只增不减——rate 算增量除以时间——得到 QPS
rate(http_server_requests_seconds_count{application="user-service"}[1m])
# 意思是：近 1 分钟内——每秒的请求增长速率 = QPS

# ③ irate()——瞬时速率（比 rate 更灵敏——但曲线毛刺多）
irate(http_server_requests_seconds_count{application="user-service"}[1m])

# ④ histogram_quantile()——从 Histogram 算分位数
# P99 延迟——99% 的请求在多少时间内完成
histogram_quantile(0.99, rate(http_server_requests_seconds_bucket[1m]))

# ⑤ sum() by()——按维度聚合
# 按应用名汇总 QPS
sum(rate(http_server_requests_seconds_count[1m])) by (application)

# QPS——每秒请求量
sum(rate(http_server_requests_seconds_count[1m])) by (application, uri)

# 错误率——5xx 错误占比
sum(rate(http_server_requests_seconds_count{status=~"5.."}[1m])) by (application)
  /
sum(rate(http_server_requests_seconds_count[1m])) by (application)
  * 100

# P99 延迟
histogram_quantile(0.99,
  sum(rate(http_server_requests_seconds_bucket[5m])) by (le, application))

# JVM 堆内存使用率
jvm_memory_used_bytes{area="heap"}
  /
jvm_memory_max_bytes{area="heap"}
  * 100

# GC 频率——每分钟 GC 次数
rate(jvm_gc_pause_seconds_count[1m])

# 或者在 Prometheus 中定义告警规则（推荐——和 Grafana 告警二选一）
# prometheus/alert_rules.yml
groups:
  - name: spring-boot
    rules:
      - alert: HighErrorRate
        expr: |
          sum(rate(http_server_requests_seconds_count{status=~"5.."}[5m])) by (application)
            /
          sum(rate(http_server_requests_seconds_count[5m])) by (application)
          * 100 > 5
        for: 5m
        labels:
          severity: critical
        annotations:
          summary: "{{ $labels.application }} 错误率超过 5%"