1.创建项目
pom.xml引入相关依赖
<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 https://maven.apache.org/xsd/maven-4.0.0.xsd"> <modelVersion>4.0.0</modelVersion> <groupId>com.olive</groupId> <artifactId>prometheus-meter-demo</artifactId> <version>0.0.1-SNAPSHOT</version> <parent> <groupId>org.springframework.boot</groupId> <artifactId>spring-boot-starter-parent</artifactId> <version>2.3.7.RELEASE</version> <relativePath /> </parent> <properties> <java.version>1.8</java.version> <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding> <project.reporting.outputEncoding>UTF-8</project.reporting.outputEncoding> <spring-boot.version>2.3.7.RELEASE</spring-boot.version> </properties> <dependencies> <dependency> <groupId>org.springframework.boot</groupId> <artifactId>spring-boot-starter-aop</artifactId> </dependency> <dependency> <groupId>org.springframework.boot</groupId> <artifactId>spring-boot-starter-web</artifactId> </dependency> <dependency> <groupId>org.springframework.boot</groupId> <artifactId>spring-boot-starter-actuator</artifactId> </dependency> <!-- Micrometer Prometheus registry --> <dependency> <groupId>io.micrometer</groupId> <artifactId>micrometer-registry-prometheus</artifactId> </dependency> </dependencies> <dependencyManagement> <dependencies> <dependency> <groupId>org.springframework.boot</groupId> <artifactId>spring-boot-dependencies</artifactId> <version>${spring-boot.version}</version> <type>pom</type> <scope>import</scope> </dependency> </dependencies> </dependencyManagement> </project>
2.自定义指标
- 方式一
直接使用micrometer核心包的类进行指标定义和注册
package com.olive.monitor; import javax.annotation.PostConstruct; import org.springframework.beans.factory.annotation.Autowired; import org.springframework.stereotype.Component; import io.micrometer.core.instrument.Counter; import io.micrometer.core.instrument.DistributionSummary; import io.micrometer.core.instrument.MeterRegistry; @Component public class NativeMetricsMontior { /** * 支付次数 */ private Counter payCount; /** * 支付金额统计 */ private DistributionSummary payAmountSum; @Autowired private MeterRegistry registry; @PostConstruct private void init() { payCount = registry.counter("pay_request_count", "payCount", "pay-count"); payAmountSum = registry.summary("pay_amount_sum", "payAmountSum", "pay-amount-sum"); } public Counter getPayCount() { return payCount; } public DistributionSummary getPayAmountSum() { return payAmountSum; } }
- 方式二
通过引入micrometer-registry-prometheus包，该包结合prometheus，对micrometer进行了封装
<dependency> <groupId>io.micrometer</groupId> <artifactId>micrometer-registry-prometheus</artifactId> </dependency>
同样定义两个metrics
package com.olive.monitor; import javax.annotation.PostConstruct; import org.springframework.beans.factory.annotation.Autowired; import org.springframework.stereotype.Component; import io.prometheus.client.CollectorRegistry; import io.prometheus.client.Counter; @Component public class PrometheusMetricsMonitor { /** * 订单发起次数 */ private Counter orderCount; /** * 金额统计 */ private Counter orderAmountSum; @Autowired private CollectorRegistry registry; @PostConstruct private void init() { orderCount = Counter.build().name("order_request_count") .help("order request count.") .labelNames("orderCount") .register(); orderAmountSum = Counter.build().name("order_amount_sum") .help("order amount sum.") .labelNames("orderAmountSum") .register(); registry.register(orderCount); registry.register(orderAmountSum); } public Counter getOrderCount() { return orderCount; } public Counter getOrderAmountSum() { return orderAmountSum; } }
prometheus 4种常用Metrics
Counter
连续增加不会减少的计数器，可以用于记录只增不减的类型，例如：网站访问人数，系统运行时间等。
对于Counter类型的指标，只包含一个inc()的方法，就是用于计数器+1.
一般而言，Counter类型的metric指标在冥冥中我们使用_total结束，如http_requests_total.
Gauge
可增可减的仪表盘，曲线图
对于这类可增可减的指标，用于反应应用的当前状态。
例如在监控主机时，主机当前空闲的内存大小，可用内存大小等等。
对于Gauge指标的对象则包含两个主要的方法inc()和dec()，用于增加和减少计数。
Histogram
主要用来统计数据的分布情况，这是一种特殊的metrics数据类型，代表的是一种近似的百分比估算数值，统计所有离散的指标数据在各个取值区段内的次数。例如：我们想统计一段时间内http请求响应小于0.005秒、小于0.01秒、小于0.025秒的数据分布情况。那么使用Histogram采集每一次http请求的时间，同时设置bucket。
Summary
Summary和Histogram非常相似，都可以统计事件发生的次数或者大小，以及其分布情况，他们都提供了对时间的计数_count以及值的汇总_sum，也都提供了可以计算统计样本分布情况的功能，不同之处在于Histogram可以通过histogram_quantile函数在服务器计算分位数。而Sumamry的分位数则是直接在客户端进行定义的。因此对于分位数的计算，Summary在通过PromQL进行查询的时候有更好的性能表现，而Histogram则会消耗更多的资源，但是相对于客户端而言Histogram消耗的资源就更少。用哪个都行，根据实际场景自由调整即可。
3. 测试
定义两个controller分别使用NativeMetricsMontior和PrometheusMetricsMonitor
package com.olive.controller; import java.util.Random; import javax.annotation.Resource; import org.springframework.web.bind.annotation.RequestMapping; import org.springframework.web.bind.annotation.RequestParam; import org.springframework.web.bind.annotation.RestController; import com.olive.monitor.NativeMetricsMontior; @RestController public class PayController { @Resource private NativeMetricsMontior monitor; @RequestMapping("/pay") public String pay(@RequestParam("amount") Double amount) throws Exception { // 统计支付次数 monitor.getPayCount().increment(); Random random = new Random(); //int amount = random.nextInt(100); if(amount==null) { amount = 0.0; } // 统计支付总金额 monitor.getPayAmountSum().record(amount); return "支付成功, 支付金额: " + amount; } }
package com.olive.controller; import java.util.Random; import javax.annotation.Resource; import org.springframework.web.bind.annotation.RequestMapping; import org.springframework.web.bind.annotation.RequestParam; import org.springframework.web.bind.annotation.RestController; import com.olive.monitor.PrometheusMetricsMonitor; @RestController public class OrderController { @Resource private PrometheusMetricsMonitor monitor; @RequestMapping("/order") public String order(@RequestParam("amount") Double amount) throws Exception { // 订单总数 monitor.getOrderCount() .labels("orderCount") .inc(); Random random = new Random(); //int amount = random.nextInt(100); if(amount==null) { amount = 0.0; } // 统计订单总金额 monitor.getOrderAmountSum() .labels("orderAmountSum") .inc(amount); return "下单成功, 订单金额: " + amount; } }
启动服务
访问http://127.0.0.1:9595/actuator/prometheus;正常看到监测数据
改变amount多次方式http://127.0.0.1:8080/order?amount=100和http://127.0.0.1:8080/pay?amount=10后；再访问http://127.0.0.1:9595/actuator/prometheus。查看监控数据
4.项目中的应用
项目中按照上面说的方式进行数据埋点监控不太现实；在spring项目中基本通过AOP进行埋点监测。比如写一个切面Aspect；这样的方式就非常友好。能在入口就做了数据埋点监测，无须在controller里进行代码编写。
package com.olive.aspect; import java.time.LocalDate; import java.util.concurrent.TimeUnit; import javax.servlet.http.HttpServletRequest; import org.aspectj.lang.ProceedingJoinPoint; import org.aspectj.lang.annotation.Around; import org.aspectj.lang.annotation.Aspect; import org.aspectj.lang.annotation.Pointcut; import org.springframework.stereotype.Component; import org.springframework.util.StringUtils; import org.springframework.web.context.request.RequestContextHolder; import org.springframework.web.context.request.ServletRequestAttributes; import io.micrometer.core.instrument.Metrics; @Aspect @Component public class PrometheusMetricsAspect { // 切入所有controller包下的请求方法 @Pointcut("execution(* com.olive.controller..*.*(..))") public void controllerPointcut() { } @Around("controllerPointcut()") public Object MetricsCollector(ProceedingJoinPoint joinPoint) throws Throwable { HttpServletRequest request = ((ServletRequestAttributes) RequestContextHolder.getRequestAttributes()).getRequest(); String userId = StringUtils.hasText(request.getParameter("userId")) ? request.getParameter("userId") : "no userId"; // 获取api url String api = request.getServletPath(); // 获取请求方法 String method = request.getMethod(); long startTs = System.currentTimeMillis(); LocalDate now = LocalDate.now(); String[] tags = new String[10]; tags[0] = "api"; tags[1] = api; tags[2] = "method"; tags[3] = method; tags[4] = "day"; tags[5] = now.toString(); tags[6] = "userId"; tags[7] = userId; String amount = StringUtils.hasText(request.getParameter("amount")) ? request.getParameter("amount") : "0.0"; tags[8] = "amount"; tags[9] = amount; // 请求次数加1 //自定义的指标名称：custom_http_request_all，指标包含数据 Metrics.counter("custom_http_request_all", tags).increment(); Object object = null; try { object = joinPoint.proceed(); } catch (Exception e) { //请求失败次数加1 Metrics.counter("custom_http_request_error", tags).increment(); throw e; } finally { long endTs = System.currentTimeMillis() - startTs; //记录请求响应时间 Metrics.timer("custom_http_request_time", tags).record(endTs, TimeUnit.MILLISECONDS); } return object; } }
编写好切面后，重启服务；访问controller的接口，同样可以进行自定义监控指标埋点
