上一篇博文中,说明了怎么引进Prometheus到asp.net core项目中,因为是Demo,所以Prometheus和Grafana都是windows版本,本地执行的,生产环境上这些服务可以根据的公司的架构,放到适合的环境内,现在这些服务都支持跨平台化和容器化。并且在上篇博客中展示的是http请求的基础信息模板,本篇博客介绍自定义Prometheusr指标类型。
Prometheus有四种指标类型:Counter(计数器)、Gauge(仪表盘)、Histogram(直方图)、Summary(摘要),如果对业务的指标进行收集展示,在项目中是侵入式编程的,如果项目使用Prometheus.net进行对接Permetheus,是通过该包中的静态方法 Metrics.CreateCounter(),Metrics.CreateGauge(),Metrics.CreateSummary(),Metrics.CreateHistogram()来创建静态指标收集器,完成对业务指标收集的。
我们先来看具体Demo。
1、Counter:计数器,只增不减
先设置个业务场景:比如做一个商城,有用户注册(/register),下订单(/order),支付(/pay),发货(/ship)四个API,代码如下:
using Microsoft.AspNetCore.Mvc;
using Microsoft.Extensions.Logging;
using PrometheusSample.Models;
using PrometheusSample.Services;
using System;
using System.Threading.Tasks;namespace PrometheusSample.Controllers
{[ApiController][Route("[controller]")]public class BusinessController : ControllerBase{private readonly ILogger<BusinessController> _logger;private readonly IOrderService _orderService;public BusinessController(ILogger<BusinessController> logger, IOrderService orderService){_orderService = orderService;_logger = logger;}/// <summary>/// 注册/// </summary>/// <param name="username">用户名</param>/// <returns></returns>[HttpPost("/register")]public async Task<IActionResult> RegisterUser([FromBody] User user){try{_logger.LogInformation("用户注册");var result = await _orderService.Register(user.UserName);if (result){return new JsonResult(new { Result = true });}else{return new JsonResult(new { Result = false });}}catch (Exception exc){_logger.LogCritical(exc, exc.Message);return new JsonResult(new { Result = false, Message = exc.Message });}}[HttpGet("/order")]public IActionResult Order(string orderno){try{_logger.LogInformation("下单"); return new JsonResult(new { Result = true });}catch (Exception exc){_logger.LogCritical(exc, exc.Message);return new JsonResult(new{Result = false,Message = exc.Message});}}[HttpGet("/pay")]public IActionResult Pay(){try{_logger.LogInformation("支付");return new JsonResult(new { Result = true });}catch (Exception exc){_logger.LogCritical(exc, exc.Message);return new JsonResult(new { Result = false, Message = exc.Message });}}[HttpGet("/ship")]public IActionResult Ship(){try{_logger.LogInformation("发货");return new JsonResult(new { Result = true });}catch (Exception exc){_logger.LogCritical(exc, exc.Message);return new JsonResult(new { Result = false, Message = exc.Message });}}}
}
上面是基本的业务Controller,为了降低依赖,我们的业务指标收集统一到一个中间件中去收集,中间件根据请求的url,和返回的数据结果数据进行业务指标数据的收集,当然也可以引入action过滤器或MediatR等中介者模式的组件来隔离业务逻辑的开发与监控数据的采集。
本例是用中间件的方式,首先定义一个静态的指标收集器:
public class MetricsHub{private static Dictionary<string, Counter> _counterDictionary = new Dictionary<string, Counter>();public Counter GetCounter(string key){if (_counterDictionary.ContainsKey(key)){return _counterDictionary[key];}else{return null;}}public void AddCounter(string key, Counter counter){_counterDictionary.Add(key, counter);}}
定义中间件BusinessMetricsMiddleware
using Microsoft.AspNetCore.Http;
using PrometheusSample.Models;
using System.IO;
using System.Threading.Tasks;namespace PrometheusSample.Middlewares
{/// <summary>/// 请求记录中间件/// </summary>public class BusinessMetricsMiddleware{private readonly RequestDelegate _next;public BusinessMetricsMiddleware(RequestDelegate next){_next = next;}public async Task InvokeAsync(HttpContext context, MetricsHub metricsHub){var originalBody = context.Response.Body;try{using (var memStream = new MemoryStream()){//从管理返回的Response中取出返回数据,根据返回值进行监控指标计数context.Response.Body = memStream;await _next(context);memStream.Position = 0;string responseBody = new StreamReader(memStream).ReadToEnd();memStream.Position = 0;await memStream.CopyToAsync(originalBody);if (metricsHub.GetCounter(context.Request.Path) != null || metricsHub.GetGauge(context.Request.Path) != null){//这里约定所有action返回值是一个APIResult类型var result = System.Text.Json.JsonSerializer.Deserialize<APIResult>(responseBody, new System.Text.Json.JsonSerializerOptions { PropertyNameCaseInsensitive = true });if (result != null && result.Result){//获取到Countervar counter = metricsHub.GetCounter(context.Request.Path);if (counter != null){//计数counter.Inc();}}}}}finally{context.Response.Body = originalBody;}}}
}
中间件中,只要action请求返回的Result为true,就会计数,这样做的前提条件是业务返回值有统一约定;但每个action返回不可能都一样的,如果有特例,可以用action过滤器或中介者模式组件来对应。
再看一下Starup中是怎么配置这个中间件的:
using Microsoft.AspNetCore.Builder;
using Microsoft.AspNetCore.Hosting;
using Microsoft.AspNetCore.Mvc;
using Microsoft.Extensions.Configuration;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Hosting;
using Microsoft.Extensions.Logging;
using Microsoft.OpenApi.Models;
using Prometheus;
using PrometheusSample.Middlewares;
using PrometheusSample.Services;
using System;
using System.Collections.Generic;
using System.Linq;
using System.Threading.Tasks;namespace PrometheusSample
{public class Startup{public Startup(IConfiguration configuration){Configuration = configuration;}public IConfiguration Configuration { get; }public void ConfigureServices(IServiceCollection services){MetricsHandle(services);services.AddScoped<IOrderService, OrderService>();services.AddControllers();services.AddSwaggerGen(c =>{c.SwaggerDoc("v1", new OpenApiInfo { Title = "PrometheusSample", Version = "v1" });});}public void Configure(IApplicationBuilder app, IWebHostEnvironment env){if (env.IsDevelopment()){app.UseDeveloperExceptionPage();app.UseSwagger();app.UseSwaggerUI(c => c.SwaggerEndpoint("/swagger/v1/swagger.json", "PrometheusSample v1"));}app.UseRouting();//http请求的中间件app.UseHttpMetrics();app.UseAuthorization();//自定义业务跟踪app.UseBusinessMetrics();app.UseEndpoints(endpoints =>{//映射监控地址为 /metricsendpoints.MapMetrics();endpoints.MapControllers();});}/// <summary>/// 处理监控事项/// </summary>/// <param name="services"></param>void MetricsHandle(IServiceCollection services){var metricsHub = new MetricsHub();//countermetricsHub.AddCounter("/register", Metrics.CreateCounter("business_register_user", "注册用户数。"));metricsHub.AddCounter("/order", Metrics.CreateCounter("business_order_total", "下单总数。"));metricsHub.AddCounter("/pay", Metrics.CreateCounter("business_pay_total", "支付总数。"));metricsHub.AddCounter("/ship", Metrics.CreateCounter("business_ship_total", "发货总数。"));services.AddSingleton(metricsHub);}}
}
MetricsHandle中,我们添加了四个action,分别对应的四个计数器,这样,当这四个url有请求,并且返回值中的result=true时,就会往对应的计数器上计数。
这样数据收集好了,现在开始在Grafana中配置显示的图表了:
订单各状态总数配置:
订单各状态30秒内数量跟踪折线
最后的运行结果是:
总结实现自定义业务计数器步骤:
1、分析业务,规划好监控跟踪指标
2、定义指标收集器
3、侵入编程(尽量在开发时分离业务实现与监控指票的收集代码)收集指标
4、开发grafana展示模板,完成展示