Windows provides a wide variety of performance counters such as CPU occupancy, memory, disk, and network usage. You can also define your own. Application Insights can show these performance counters if your application is running under IIS on an on-premises host or virtual machine to which you have administrative access. The charts indicate the resources available to your live application, and can help to identify unbalanced load between server instances.
Performance counters appear in the Servers blade, which includes a table that segments by server instance.
(Performance counters aren't available for Azure Web Apps. But you can send Azure Diagnostics to Application Insights.)
The Servers blade shows a default set of performance counters.
To see other counters, either edit the charts on the Servers blade, or open a new Metrics Explorer blade and add new charts.
The available counters are listed as metrics when you edit a chart.
To see all your most useful charts in one place, create a dashboard and pin them to it.
If the performance counter you want isn't shown in the list of metrics, that's because the Application Insights SDK isn't collecting it in your web server. You can configure it to do so.
Find out what counters are available in your server by using this PowerShell command at the server:
Get-Counter -ListSet *
- If you added Application Insights to your app during development, edit ApplicationInsights.config in your project, and then re-deploy it to your servers.
- If you used Status Monitor to instrument a web app at runtime, find ApplicationInsights.config in the root directory of the app in IIS. Update it there in each server instance.
- Edit the performance collector directive:
<Add Type="Microsoft.ApplicationInsights.Extensibility.PerfCounterCollector.PerformanceCollectorModule, Microsoft.AI.PerfCounterCollector"> <Counters> <Add PerformanceCounter="\Objects\Processes"/> <Add PerformanceCounter="\Sales(photo)\# Items Sold" ReportAs="Photo sales"/> </Counters> </Add>
You can capture both standard counters and those you have implemented yourself.
\Objects\Processes is an example of a standard counter, available on all Windows systems.
\Sales(photo)\# Items Sold is an example of a custom counter that might be implemented in a web service.
The format is
\Category(instance)\Counter", or for categories that don't have instances, just
ReportAs is required for counter names that do not match
[a-zA-Z()/-_ \.]+ - that is, they contain characters that are not in the following sets: letters, round brackets, forward slash, hyphen, underscore, space, dot.
If you specify an instance, it will be collected as a dimension "CounterInstanceName" of the reported metric.
Collecting performance counters in code
To collect system performance counters and send them to Application Insights, you can adapt the snippet below:
var perfCollectorModule = new PerformanceCollectorModule(); perfCollectorModule.Counters.Add(new PerformanceCounterCollectionRequest( @"\.NET CLR Memory([replace-with-application-process-name])\# GC Handles", "GC Handles"))); perfCollectorModule.Initialize(TelemetryConfiguration.Active);
Or you can do the same thing with custom metrics you created:
var perfCollectorModule = new PerformanceCollectorModule(); perfCollectorModule.Counters.Add(new PerformanceCounterCollectionRequest( @"\Sales(photo)\# Items Sold", "Photo sales")); perfCollectorModule.Initialize(TelemetryConfiguration.Active);
Performance counters in Analytics
You can search and display performance counter reports in Analytics.
The performanceCounters schema exposes the
counter name, and
instance name of each performance counter. In the telemetry for each application, you’ll see only the counters for that application. For example, to see what counters are available:
('Instance' here refers to the performance counter instance, not the role or server machine instance. The performance counter instance name typically segments counters such as processor time by the name of the process or application.)
To get a chart of available memory over the recent period:
Like other telemetry, performanceCounters also has a column
cloud_RoleInstance that indicates the identity of the host server instance on which your app is running. For example, to compare the performance of your app on the different machines:
ASP.NET and Application Insights counts
What's the difference between the Exception rate and Exceptions metrics?
- Exception rate is a system performance counter. The CLR counts all the handled and unhandled exceptions that are thrown, and divides the total in a sampling interval by the length of the interval. The Application Insights SDK collects this result and sends it to the portal.
- Exceptions is a count of the TrackException reports received by the portal in the sampling interval of the chart. It includes only the handled exceptions where you have written TrackException calls in your code, and doesn't include all unhandled exceptions.
Like other metrics, you can set an alert to warn you if a performance counter goes outside a limit you specify. Open the Alerts blade and click Add Alert.