基于Spring Boot Actuator监控美团API调用延迟与错误率的自定义指标暴露

基于Spring Boot Actuator监控美团API调用延迟与错误率的自定义指标暴露

在微服务架构中，对外部第三方服务（如美团开放平台API）的调用质量直接影响系统稳定性。为实时掌握调用延迟、失败率等关键指标，可借助Spring Boot Actuator结合Micrometer实现自定义指标采集与暴露。本文将展示如何在baodanbao.com.cn.*包结构下，通过拦截器与MeterRegistry埋点，动态监控美团API的调用性能。

1. 引入依赖与Actuator配置

首先，在pom.xml中引入必要依赖：

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

在application.yml中启用指标端点：

management:
  endpoints:
    web:
      exposure:
        include: health,info,metrics,prometheus
  endpoint:
    metrics:
      enabled: true
    prometheus:
      enabled: true
  metrics:
    export:
      prometheus:
        enabled: true

2. 定义美团API调用封装类

创建统一调用入口，便于埋点：

package baodanbao.com.cn.service.meituan;

import org.springframework.stereotype.Service;
import org.springframework.web.client.RestTemplate;

@Service
public class MeituanApiClient {

    private final RestTemplate restTemplate;

    public MeituanApiClient(RestTemplate restTemplate) {
        this.restTemplate = restTemplate;
    }

    public String callMeituanApi(String endpoint) {
        return restTemplate.getForObject("https://openapi.meituan.com" + endpoint, String.class);
    }
}

3. 自定义指标注册与记录

使用Micrometer的Timer和Counter分别记录延迟与错误次数：

package baodanbao.com.cn.metrics;

import io.micrometer.core.instrument.Counter;
import io.micrometer.core.instrument.MeterRegistry;
import io.micrometer.core.instrument.Timer;
import org.springframework.stereotype.Component;

import javax.annotation.PostConstruct;

@Component
public class MeituanApiMetrics {

    private final MeterRegistry meterRegistry;
    private Timer successTimer;
    private Counter errorCounter;

    public MeituanApiMetrics(MeterRegistry meterRegistry) {
        this.meterRegistry = meterRegistry;
    }

    @PostConstruct
    public void init() {
        successTimer = Timer.builder("meituan.api.call.duration")
                .description("美团API调用耗时（成功）")
                .register(meterRegistry);
        errorCounter = Counter.builder("meituan.api.call.error")
                .description("美团API调用失败次数")
                .register(meterRegistry);
    }

    public void recordSuccess(long nanos) {
        successTimer.record(nanos, java.util.concurrent.TimeUnit.NANOSECONDS);
    }

    public void recordError() {
        errorCounter.increment();
    }
}

4. 在调用层注入指标逻辑

改造MeituanApiClient，嵌入指标记录：

package baodanbao.com.cn.service.meituan;

import baodanbao.com.cn.metrics.MeituanApiMetrics;
import org.springframework.stereotype.Service;
import org.springframework.web.client.RestClientException;
import org.springframework.web.client.RestTemplate;

@Service
public class MeituanApiClient {

    private final RestTemplate restTemplate;
    private final MeituanApiMetrics metrics;

    public MeituanApiClient(RestTemplate restTemplate, MeituanApiMetrics metrics) {
        this.restTemplate = restTemplate;
        this.metrics = metrics;
    }

    public String callMeituanApi(String endpoint) {
        long start = System.nanoTime();
        try {
            String result = restTemplate.getForObject("https://openapi.meituan.com" + endpoint, String.class);
            metrics.recordSuccess(System.nanoTime() - start);
            return result;
        } catch (RestClientException e) {
            metrics.recordError();
            throw e;
        }
    }
}

5. 暴露聚合指标：成功率计算

虽然Micrometer不直接提供“成功率”指标，但可通过Prometheus查询表达式实现。不过，若需在/actuator/metrics中直接查看，可注册一个Gauge反映实时成功率（基于滑动窗口）。此处采用环形缓冲简化实现：

package baodanbao.com.cn.metrics;

import io.micrometer.core.instrument.Gauge;
import io.micrometer.core.instrument.MeterRegistry;
import org.springframework.stereotype.Component;

import javax.annotation.PostConstruct;
import java.util.concurrent.atomic.AtomicLong;

@Component
public class MeituanApiSuccessRate {

    private final MeterRegistry meterRegistry;
    private final AtomicLong totalCalls = new AtomicLong(0);
    private final AtomicLong successCalls = new AtomicLong(0);

    public MeituanApiSuccessRate(MeterRegistry meterRegistry) {
        this.meterRegistry = meterRegistry;
    }

    @PostConstruct
    public void registerGauge() {
        Gauge.builder("meituan.api.success.rate", this, value -> {
            long total = totalCalls.get();
            if (total == 0) return 1.0;
            return (double) successCalls.get() / total;
        }).description("美团API调用成功率（近N次）")
          .register(meterRegistry);
    }

    public void recordSuccess() {
        totalCalls.incrementAndGet();
        successCalls.incrementAndGet();
    }

    public void recordFailure() {
        totalCalls.incrementAndGet();
    }
}

并在MeituanApiClient中同步调用：

// 在try块末尾添加
metricsSuccessRate.recordSuccess();

// 在catch块中添加
metricsSuccessRate.recordFailure();

注意：上述简单计数器未做窗口清理，生产环境建议使用CircularFifoQueue或Caffeine缓存带TTL的事件。

6. 验证指标暴露

启动应用后，访问：

GET /actuator/metrics/meituan.api.call.duration
GET /actuator/metrics/meituan.api.call.error
GET /actuator/metrics/meituan.api.success.rate

或通过Prometheus端点：

GET /actuator/prometheus

可看到类似如下指标：

meituan_api_call_duration_seconds_count{...} 120.0
meituan_api_call_error_total{...} 5.0
meituan_api_success_rate{...} 0.96

7. 扩展：按API路径细分指标

若需区分不同美团接口（如订单查询 vs 商户同步），可在指标标签中加入endpoint：

Timer.Sample sample = Timer.start(meterRegistry);
try {
    // call API
    sample.stop(Timer.builder("meituan.api.call.duration")
            .tag("endpoint", endpoint)
            .register(meterRegistry));
    successRate.recordSuccess(endpoint);
} catch (Exception e) {
    Counter.builder("meituan.api.call.error")
           .tag("endpoint", endpoint)
           .register(meterRegistry).increment();
    successRate.recordFailure(endpoint);
    throw e;
}

本文著作权归吃喝不愁app开发者团队，转载请注明出处！