Using instrumentation libraries

当你开发应用时，可能会使用第三方库和框架来加快开发进度。如果你随后使用 OpenTelemetry 对应用进行插桩，你可能希望避免额外花时间为所用的第三方库和框架手动添加链路、日志和指标。

许多库和框架已经原生支持 OpenTelemetry，或者通过 OpenTelemetry 的插桩获得支持， 因此它们能够生成可导出到可观测性后端的遥测数据。

如果你正在为使用第三方库或框架的应用或服务进行插桩， 请按照以下说明学习如何为你的依赖项使用原生插桩库和插桩库。

使用原生插桩库

如果某个库默认就支持 OpenTelemetry，你只需在应用中添加并配置 OpenTelemetry SDK， 就可以获取该库发出的链路、指标和日志。

该库可能需要一些额外的插桩配置。请查阅该库的文档以了解更多信息。

• Azure SDK Instrumentation
• MassTransit .NET client
• NServiceBus OpenTelemetry Integration
• ThrottlingTroll

Use Instrumentation Libraries

If a library does not come with OpenTelemetry out of the box, you can use instrumentation libraries in order to generate telemetry data for a library or framework.

For example, the instrumentation library for ASP.NET Core will automatically create spans and metrics based on the inbound HTTP requests.

Setup

Each instrumentation library is a NuGet package, and installing them is typically done like so:

dotnet add package OpenTelemetry.Instrumentation.{library-name-or-type}

It is typically then registered at application startup time, such as when creating a TracerProvider.

Note on Versioning

The Semantic Conventions (Standards) for attribute names are not currently stable therefore the instrumentation library is currently not in a released state. That doesn’t mean that the functionality itself is not stable, only that the names of some of the attributes may change in the future, some may be added, some may be removed. This means that you need to use the --prerelease flag, or install a specific version of the package

Example with ASP.NET Core and HttpClient

As an example, here’s how you can instrument inbound and output requests from an ASP.NET Core app.

First, get the appropriate packages of OpenTelemetry Core:

dotnet add package OpenTelemetry
dotnet add package OpenTelemetry.Extensions.Hosting
dotnet add package OpenTelemetry.Exporter.Console

Then you can install the instrumentation libraries:

dotnet add package OpenTelemetry.Instrumentation.AspNetCore --prerelease
dotnet add package OpenTelemetry.Instrumentation.Http --prerelease

Next, configure each instrumentation library at startup and use them!

using OpenTelemetry.Resources;
using OpenTelemetry.Trace;


var builder = WebApplication.CreateBuilder(args);

builder.Services.AddOpenTelemetry()
  .WithTracing(b =>
  {
      b
      .AddHttpClientInstrumentation()
      .AddAspNetCoreInstrumentation();
  });

var app = builder.Build();

var httpClient = new HttpClient();

app.MapGet("/hello", async () =>
{
    var html = await httpClient.GetStringAsync("https://example.com/");
    if (string.IsNullOrWhiteSpace(html))
    {
        return "Hello, World!";
    }
    else
    {
        return "Hello, World!";
    }
});

app.Run();

When you run this code and access the /hello endpoint, the instrumentation libraries will:

• Start a new trace
• Generate a span representing the request made to the endpoint
• Generate a child span representing the HTTP GET made to https://example.com/

If you add more instrumentation libraries, then you get more spans for each of those.

Available instrumentation libraries

A full list of instrumentation libraries produced by OpenTelemetry is available from the opentelemetry-dotnet repository.

You can also find more instrumentations available in the registry.

Next steps

After you have set up instrumentation libraries, you may want to add your own instrumentation to your code, to collect custom telemetry data.

If you are using .NET Framework 4.x instead of modern .NET, refer to the .NET Framework docs to configure OpenTelemetry and instrumentation libraries on .NET Framework.

You’ll also want to configure an appropriate exporter to export your telemetry data to one or more telemetry backends.

You can also check the automatic instrumentation for .NET, which is currently in beta.