Learn to use Visual Studio, Visual Studio Online, Application Insights and Team Foundation Server to decrease rework, increase transparency into your application and increase the rate at which you can ship high quality software throughout the application lifecycle
Application Insights has only been available for a month; so it doesn’t come as any surprise that most of the questions and issues have been on the setup process and many of those question has been around the setup of the Microsoft Monitoring Agent. This document should help with the issues we have seen to date.
To trouble shoot these issues and more start by looking at the events in the Operations Manager event log(Directions for doing this below). Pay particular attention to APM Agent Event 4003, which is the Start of Microsoft Monitoring agent.
Luckily the solution is easy: either turn off these utilities or run the application you want to monitor under an Application Pool that does not have Tracelog/another profiler/agent enabled. Unfortunately there isn’t an easy way to see if this causing your problem…So either look into the event log (see above) or simply create a new Application Pool and run your application in this Application Pool.
3. If you have the same time range but are seeing different data sets for the same application it is likely you have you Visual Studio Online profile to use “high contrast”…We are working on solving these issues for Application Insights in our next sprint!
4. The application being monitored isn’t seeing any traffic. The application needs to be getting traffic as we won’t produce data if the application pool is not running.
5. (related to above) The traffic is going to the wrong web application. Remember Visual Studio defaults to running applications in IIS Express (which we don’t monitor) and your requests can easily be redirected! To work around this try generating load using Localhost.
6. If the IIS Application Pool recycling threshold was set too low and IIS gets into a loop of recycling it the Application pool will be disabled. As stated above the application needs to be getting traffic – we won’t produce data if the application pool is down. During debugging sessions this is a pretty common issue!
7. For services running asynchronously remember exception data will often be delayed.
8. If Microsoft Monitoring Agent is in a compromised or non responsive state it will stop sending data and may need to be restarted. To restart Application Insight services please restart the following local services (see directions below):
3. Expand the Event Viewer folder, the Windows Logs folder, the Applications and Services Logs folder then click on the Operations Manager icon.
Since most trouble shooting of the Microsoft Monitoring Agent will involve looking at the event log: enclosed below is a list of both Application Insights and other services Event IDs to help you trouble-shoot issues you may encounter.
.NET app monitoring
Application was removed from monitoring (IIS Recycle/Restart required)
Application was added to monitoring (IIS Recycle/Restart required)
Application settings have been updated (IIS Recycle/Restart required)
Application Exception monitoring settings have been changed (IIS Recycle/Restart required)
Web application being monitored has conflicts in configuration. Monitoring is disabled
Configuration successfully applied
Configuration contains errors. Application monitoring failed to start
Configuration successfully applied. Some non-critical conflicts detected
Configuration processing has started
Configuration processing is completed
System Center Management APM service is running
Unrecoverable error during processing of configuration
Write Action configuration infrastructure has stopped
Different errors from the configuration processing system
Microsoft Monitoring Agent started successfully
Health Service Script
IIS was successfully restarted
Microsoft Monitoring Agent is not running
Namespace collection disabled (not strictly a failure; effect of optimization – but can prevent you from seeing expected data)
Configuration files have been modified and reloaded
Performance counters are being collected
Shut down of Microsoft Monitoring Agent has occurred
I had a problem where the data wouldn't appear in AI and it turned out to be related to network connectivity issues related to a proxy server. The Microsoft Monitoring Agent service runs as Local System user, and that user hadn't been configured to use our site's proxy server. Consequently it didn't have access to the Internet.
The error in the event log is:
Event ID: 2130
Description: Health Service HTTP module in management group "AppDiagnostics" timed out connecting to url (eus-amt-wt1.tofino.windowsazure.com/.../TelemetryData).
To get that user to use the proxy server settings of the current interactive user, I used this command:
> netsh winhttp import proxy ie
There are probably other ways to set up the user.
I don't know if there's a different way to get the MMA service to use specific proxy settings. There are a bunch of config files in the MMA install folder - can we use the standard <system.net><defaultProxy> for .NET applications, for example?