How can I determine the root cause of CLR crashes?

[pclark]

Updated on 7/30/18

Question

We see CLR crashes happening a few times a day. How can I use AppDynamics to determine the root cause of the crashes?

Answer

One issue usually causes multiple CLR crashes. Here are some places to look to help you find the root cause:

1. When a CLR crashes, the .NET Machine Agent raises an event in the Controller. Start by looking at the details captured in the CLR crash event by clicking on the Server Crashes link in the Application Dashboard. You will see information about the crash event, like the process ID and windows event log id. Double-click on the CLR Crash Type to view additional information that can help you troubleshoot the root cause of the crash. For additional instructions on reviewing the crash event, see: Analyze and Respond to CLR Crashes

2. Look at any errors/exceptions happening around the same time as the crash that could be related to it. Transaction Analysis is a great place to get a combined view (see below). If the crash occurred before we could report the captured exception, you may not see anything there.

3. Look at Machine snapshots from that time on the machine where the crash occurred. This will rule out any physical resource pressures.

Related Resources

• List of crash events the .NET machine agent monitors
• Monitor CLRs