Spiffy.Monitoring.Prometheus 1.2.0

Overview

A structured logging framework for .NET that supports log aggregation, e.g. Splunk.

Handled over 789,000,000,000 production requests (and counting!)

Status

Package	Latest Release
Spiffy.Monitoring
Spiffy.Monitoring.Aws
Spiffy.Monitoring.NLog
Spiffy.Monitoring.Prometheus
Spiffy.Monitoring.Splunk

Setup

PM> Install-Package Spiffy.Monitoring

Built-In Logging Providers

Spiffy.Monitoring includes "built-in" logging mechanisms (Trace and Console).

There is no default logging behavior, you must initialize provider(s) by calling Spiffy.Monitoring.Configuration.Initialize.

Until initialized, any published EventContext will not be observable, so it is recommended that initialization be as early as possible when your application is starting (i.e. in the entry point).

Example

  Configuration.Initialize(spiffy => { spiffy.Providers.Console(); });

Extended Providers

For extended functionality, you'll need to install a "provider package".

NOTE: the provider package need only be installed for your application's entry point assembly, it need not be installed in library packages.

NLog Provider

PM> Install-Package Spiffy.Monitoring.NLog

Example

    static void Main() {
        // this should be the first line of your application
        Spiffy.Monitoring.Configuration.Initialize(spiffy => {
            spiffy.Providers
                .NLog(nlog => nlog.Targets(t => t.File()));
        });
    }

Multiple Providers

Multiple providers can be provied, for example, this application uses both Console (built-in), as well as File (NLog)

Example

    Spiffy.Monitoring.Configuration.Initialize(spiffy => {
        spiffy.Providers
            .Console()
            .NLog(nlog => nlog.Targets(t => t.File()));
    });

Log

Example Program

        // key-value-pairs set here appear in every event message
        GlobalEventContext.Instance
            .Set("Application", "MyApplication");

        using (var context = new EventContext()) {
            context["Key"] = "Value";

            using (context.Time("LongRunning")) {
                DoSomethingLongRunning();
            }

            try {
                DoSomethingDangerous();
            }
            catch (Exception ex) {
                context.IncludeException(ex);
            }
        }

Normal Entry

[2014-06-13 00:05:17.634Z] Application=MyApplication Level=Info Component=Program Operation=Main TimeElapsed=1004.2 Key=Value TimeElapsed_LongRunning=1000.2

Exception Entry

[2014-06-13 00:12:52.038Z] Application=MyApplication Level=Error Component=Program Operation=Main TimeElapsed=1027.0 Key=Value ErrorReason="An exception has ocurred" Exception_Type=ApplicationException Exception_Message="you were unlucky!" Exception_StackTrace=" at TestConsoleApp.Program.DoSomethingDangerous() in c:\src\git\github\chris-peterson\Spiffy\src\Tests\TestConsoleApp\Program.cs:line 47 at TestConsoleApp.Program.Main() in c:\src\git\github\chris-peterson\Spiffy\src\Tests\TestConsoleApp\Program.cs:line 29" InnermostException_Type=NullReferenceException InnermostException_Message="Object reference not set to an instance of an object." InnermostException_StackTrace={null} Exception="See Exception_* and InnermostException_* for more details" TimeElapsed_LongRunning=1000.0

Hosting Frameworks

Spiffy.Monitoring is designed to be easy to use in any context.

The most basic usage is to instrument a specific method. This can be achieved by "newing up" an EventContext. This usage mode results in Component being set to the containing code's class name, and Operation is set to the containing code's method name

There are times when you may want to instrument something that's not a specific method. One such example is an API -- in this context, you might want to have 1 log event per request. Consider setting Component to be the controller name, and Operation the action name. To acheive this, add middleware that calls EventContext.Initialize with the desired labels.