As usual, you can obtain a reference to ILogger<T> from the DI framework and use it in your views, like this:
@inject ILogger<MyView> Logger
But there is also another built-in mechanism, the DiagnosticSource class, and property, which is declared in the RazorPage base class. By calling its Write method, you can write custom messages to a diagnostics framework. These messages can be any .NET object, even an anonymous one, and there is no need to worry about its serialization. Have a look at the following code snippet:
@{ DiagnosticSource.Write("MyDiagnostic", new { data = "A diagnostic" }); }
What happens with this diagnostic message is actually somewhat configurable. First, let's add the Microsoft.Extensions.DiagnosticAdapter NuGet package, and then create a custom listener for the events generated for this diagnostic source, like this:
public class DiagnosticListener { [DiagnosticName("MyDiagnostic")] public virtual void OnDiagnostic(string data) { //do something with data } }
We can add as many listeners as we want, targeting different event names. The actual method name does not matter, as long as it has a [DiagnosticName] attribute applied to it that matches an event name. We need to register and hook it to the .NET Core framework, in the Configure method, by adding a reference to the DiagnosticListener service so that we can interact with it, like this:
public void Configure(IApplicationBuilder app, DiagnosticListener diagnosticListener) { var listener = new DiagnosticListener(); diagnosticListener.SubscribeWithAdapter(listener);
//rest goes here }
Notice that the name in the [DiagnosticName] attribute and DiagnosticSource.Write call match, and also, the name, data, of the anonymous type in the Write call matches the parameter name (and type) of the OnDiagnostic method.
Built-in .NET Core classes produce diagnostics for the following:
Hopefully, the names should be self-explanatory. Why would you use this mechanism over the ILogger-based one? This one makes it very easy to add listeners to a diagnostic source, with strongly typed methods. I will talk more about the differences between the two in Chapter 12, Logging, Tracing, and Diagnostics.