configuring-opentelemetry-dotnet

在 .NET 中配置 OpenTelemetry

适用场景

• 为 ASP.NET Core 应用添加分布式追踪
• 设置 OpenTelemetry 导出器（OTLP 是主要协议；Jaeger 原生支持 OTLP；Prometheus 的 OTLP 数据接收需要显式启用）
• 为业务操作创建自定义指标或追踪跨度
• 排查跨服务的分布式追踪上下文传播问题

不适用场景

• 用户只需要应用级日志记录（使用 ILogger、Serilog）
• 用户直接使用 Application Insights SDK（API 不同）
• 用户需要使用商业供应商专有 SDK 的 APM

输入

输入	必需	描述
ASP.NET Core 项目	是	要进行插桩的应用
可观测性后端	否	数据导出目标：OTLP 收集器、Aspire 仪表板、Jaeger（原生支持 OTLP）

工作流

步骤 1：安装正确的程序包

OpenTelemetry NuGet 程序包种类繁多。请仅安装以下程序包：

# Core SDK + ASP.NET Core instrumentation + logging integration
dotnet add package OpenTelemetry.Extensions.Hosting
dotnet add package OpenTelemetry.Instrumentation.AspNetCore
dotnet add package OpenTelemetry.Instrumentation.Http

# Exporter
dotnet add package OpenTelemetry.Exporter.OpenTelemetryProtocol  # OTLP exporter for traces, metrics, AND logs

# Optional — dev/local debugging only (do NOT include in production deployments)
# dotnet add package OpenTelemetry.Exporter.Console

不要只安装 OpenTelemetry——还需要 OpenTelemetry.Extensions.Hosting，才能正确集成 DI。

可选：其他自动插桩程序包

只安装与你的应用所用库相匹配的程序包：

dotnet add package OpenTelemetry.Instrumentation.SqlClient           # SQL Server queries
dotnet add package OpenTelemetry.Instrumentation.EntityFrameworkCore  # EF Core
dotnet add package OpenTelemetry.Instrumentation.GrpcNetClient       # gRPC calls
dotnet add package OpenTelemetry.Instrumentation.Runtime             # GC, thread pool metrics

步骤 2：在 Program.cs 中配置所有信号

using OpenTelemetry.Resources;
using OpenTelemetry.Trace;
using OpenTelemetry.Metrics;
using OpenTelemetry.Logs;

var builder = WebApplication.CreateBuilder(args);

builder.Services.AddOpenTelemetry()
    .ConfigureResource(resource => resource
        .AddService(serviceName: builder.Environment.ApplicationName))
    .WithTracing(tracing => tracing
        .AddAspNetCoreInstrumentation(options =>
        {
            // Filter out health check endpoints from traces
            options.Filter = httpContext =>
                !httpContext.Request.Path.StartsWithSegments("/healthz");
        })
        .AddHttpClientInstrumentation(options =>
        {
            options.RecordException = true;
        })
        // Optional: add SQL instrumentation if using SqlClient directly
        // .AddSqlClientInstrumentation(options =>
        // {
        //     options.SetDbStatementForText = true;
        //     options.RecordException = true;
        // })
        // Custom activity sources (must match ActivitySource names in your code)
        .AddSource("MyApp.Orders")
        .AddSource("MyApp.Payments")
        .AddSource("MyApp.Messaging"))
    .WithMetrics(metrics => metrics
        .AddAspNetCoreInstrumentation()
        .AddHttpClientInstrumentation()
        // Optional: .AddRuntimeInstrumentation() for GC and thread pool metrics
        //   (requires OpenTelemetry.Instrumentation.Runtime package)
        // Custom meters (must match Meter names in your code)
        .AddMeter("MyApp.Metrics"))
    .WithLogging(logging =>
    {
        logging.IncludeScopes = true;
        // logging.IncludeFormattedMessage = true;  // Enable if you need the formatted message string in log exports
    })
    // Single OTLP exporter for all signals — reads OTEL_EXPORTER_OTLP_ENDPOINT
    // env var (defaults to http://localhost:4317). Override via environment variable
    // or appsettings.json configuration.
    .UseOtlpExporter();

步骤 3：了解日志与追踪的关联

步骤 2 中的 .WithLogging() 调用会将 ILogger 与 OpenTelemetry 集成：

• 每条日志记录都会自动包含 TraceId 和 SpanId，以便与追踪关联
• .ConfigureResource() 中配置的服务资源会自动传播到日志
• UseOtlpExporter() 同样适用于日志和指标，不仅限于追踪
• 无需额外程序包或单独调用 SetResourceBuilder

步骤 4：为业务操作创建自定义跨度（Activities）

using System.Diagnostics;
using Microsoft.Extensions.Logging;

public class OrderService
{
    // Create an ActivitySource matching what you registered in Step 2
    private static readonly ActivitySource ActivitySource = new("MyApp.Orders");
    private readonly ILogger<OrderService> _logger;

    public OrderService(ILogger<OrderService> logger) => _logger = logger;

    public async Task<Order> ProcessOrderAsync(CreateOrderRequest request)
    {
        // Start a new span
        using var activity = ActivitySource.StartActivity("ProcessOrder");

        // Add attributes (tags) to the span
        activity?.SetTag("order.customer_id", request.CustomerId);
        activity?.SetTag("order.item_count", request.Items.Count);

        try
        {
            // Child span for validation
            using (var validationActivity = ActivitySource.StartActivity("ValidateOrder"))
            {
                await ValidateOrderAsync(request);
                validationActivity?.SetTag("validation.result", "passed");
            }

            // Child span for payment
            using (var paymentActivity = ActivitySource.StartActivity("ProcessPayment",
                ActivityKind.Client))  // Client = outgoing call
            {
                paymentActivity?.SetTag("payment.method", request.PaymentMethod);
                await ProcessPaymentAsync(request);
            }

            var order = new Order { Id = Guid.NewGuid(), CustomerId = request.CustomerId, Status = "Completed" };

            activity?.SetTag("order.status", "completed");
            activity?.SetStatus(ActivityStatusCode.Ok);

            return order;
        }
        catch (Exception ex)
        {
            activity?.SetStatus(ActivityStatusCode.Error, ex.Message);
            // Log via ILogger — OpenTelemetry captures this with trace correlation.
            // Prefer logging over activity.RecordException() as OTel is deprecating
            // span events for exception recording in favor of log-based exceptions.
            _logger.LogError(ex, "Order processing failed for customer {CustomerId}", request.CustomerId);
            throw;
        }
    }
}

重要：ActivitySource 名称必须与配置中的 AddSource("...") 相匹配。 不匹配的来源会被静默忽略——这是最常见的调试问题。

步骤 5：创建自定义指标

使用 IMeterFactory（通过 DI 注入）创建计量器，以确保生命周期管理得当且便于测试。

using System.Diagnostics;
using System.Diagnostics.Metrics;

public class OrderMetrics
{
    private readonly Counter<long> _ordersProcessed;
    private readonly Histogram<double> _orderProcessingDuration;
    private readonly UpDownCounter<int> _activeOrders;

    public OrderMetrics(IMeterFactory meterFactory)
    {
        // Meter name must match AddMeter("...") in configuration
        var meter = meterFactory.Create("MyApp.Metrics");

        // Counter — use for things that only go up
        _ordersProcessed = meter.CreateCounter<long>(
            "orders.processed", "orders", "Total orders successfully processed");

        // Histogram — use for measuring distributions (latency, sizes)
        _orderProcessingDuration = meter.CreateHistogram<double>(
            "orders.processing_duration", "ms", "Time to process an order");

        // UpDownCounter — use for things that go up AND down
        _activeOrders = meter.CreateUpDownCounter<int>(
            "orders.active", "orders", "Currently processing orders");
    }

    public void RecordOrderProcessed(string region, double durationMs)
    {
        // Tags enable dimensional filtering (by region, status, etc.)
        var tags = new TagList
        {
            { "region", region },
            { "order.type", "standard" }
        };

        _ordersProcessed.Add(1, tags);
        _orderProcessingDuration.Record(durationMs, tags);
    }
}

在 DI 中注册 OrderMetrics：

builder.Services.AddSingleton<OrderMetrics>();

步骤 6：为分布式场景配置上下文传播

使用 AddHttpClientInstrumentation() 时，HTTP 调用会自动传播追踪上下文。对于非 HTTP 场景：

using System;
using System.Collections.Generic;
using System.Diagnostics;
using OpenTelemetry.Context.Propagation;

// ActivitySource should be static — register via .AddSource("MyApp.Messaging") in Step 2
private static readonly ActivitySource MessageSource = new("MyApp.Messaging");

// Manual context propagation (e.g., across message queues)
// On the SENDING side:
var propagator = Propagators.DefaultTextMapPropagator;
var activityContext = Activity.Current?.Context ?? default;
var context = new PropagationContext(activityContext, Baggage.Current);
var carrier = new Dictionary<string, string>();

propagator.Inject(context, carrier, (dict, key, value) => dict[key] = value);
// Send carrier dictionary as message headers

// On the RECEIVING side:
var parentContext = propagator.Extract(default, carrier,
    (dict, key) => dict.TryGetValue(key, out var value) ? new[] { value } : Array.Empty<string>());

Baggage.Current = parentContext.Baggage;
using var activity = MessageSource.StartActivity("ProcessMessage",
    ActivityKind.Consumer,
    parentContext.ActivityContext);  // Links to parent trace!

验证

• 追踪数据出现在可观测性后端（Jaeger、Aspire 仪表板等）
• HTTP 请求会自动创建包含正确动词、URL 和状态码的跨度
• 自定义 ActivitySource 名称与 AddSource() 注册项相匹配
• 自定义 Meter 名称与 AddMeter() 注册项相匹配
• 日志包含 TraceId 和 SpanId，以便进行关联
• 健康检查端点已从追踪中筛除
• 错误跨度中包含异常详情

常见问题

问题	解决方案
ActivitySource.StartActivity 返回 null	来源名称与任何 AddSource() 都不匹配——名称必须完全一致
导出器中没有追踪数据	检查 OTLP 端点：gRPC 使用端口 4317，HTTP 使用端口 4318
缺少 HTTP 客户端跨度	确保已注册 AddHttpClientInstrumentation()；它适用于 IHttpClientFactory/DI 和 new HttpClient()（为管理生命周期，请使用 IHttpClientFactory）
标签基数过高	不要将用户 ID、请求 ID 或 UUID 用作指标标签——这会导致存储量激增
OTLP 的 gRPC 与 HTTP 不匹配	默认使用 gRPC（端口 4317）；如果收集器只接受 HTTP，请设置 OtlpExportProtocol.HttpProtobuf
Meter / ActivitySource 生命周期	ActivitySource 应为静态；通过 DI 中的 IMeterFactory 创建 Meter（不要使用 new Meter()），以确保生命周期管理得当且便于测试