Monitor, Monitor, Monitor

I once heard someone say that "the great thing about Dynamics CRM is that it just looks after itself" Whilst CRM2013 is certainly very good at performing maintenance tasks automatically, if you have a customised system it is important to Monitor, Monitor, Monitor! There are some advanced ways of setting up monitoring using tools such as System Center but just some regular simple monitoring tasks will go a long way for very little investment on your part:

1) Plugin Execution Monitoring

There is a super little entity called 'Plugin-in Type Statistics' that often seems to be overlooked in the long list of advanced find entities. This entity is invaluable for tracing down issues before they cause problems for your users and as defined by the SDK it is "used by the Microsoft Dynamics CRM 2011 and Microsoft Dynamics CRM Online platforms to record execution statistics for plug-ins registered in the sandbox (isolation mode)."

The key here is that it only records statistics for your sandboxed plugins. Unless there is a good reason not to (security access etc.) I would recommend that all of your plugins be registered in sandbox isolation. Of course Dynamics CRM online only allows sandboxed plugins anyway so you don't want to put up barriers not to move to the cloud.

To monitor this you can use advanced to show a sorted list by execution time or failure count descending:

If you spot any issues you can then proactively investigate them before they become a problem. In the screen shot above there are a few plugins that are taking more than 1000ms (1 second) to execute, but their execution count is low. I look for plugins that have a high execution count and high execution time, or those that have a high failure percent.

2) Workflow & Asynchronous Job Execution Monitoring

We all know workflows often can start failing for various reasons. Because of their asynchronous nature these failures can go unnoticed by users until it's too late and you have thousands of issues to correct. To proactively monitor this you can create a view (and even add to a dashboard) of System Jobs filtered by Status = Failed or Waiting and where the Message contains data. The Message attribute contains the full error description and stack trace, but the Friendly message just contains the information that is displayed at the top of the workflow form in the notification box.

3) Client Latency & Bandwidth Monitoring

Now that you've got the server-side under control you should also look at the client connectivity of your users. There is a special diagnostics hidden page that can be accessed by using a URL of the format:

http://<YourCRMServerURL>/tools/diagnostics/diag.aspx

As described by the implementation guide topic, "Microsoft Dynamics CRM is designed to work best over networks that have the following elements:

  • Bandwidth greater than 50 KB/sec
  • Latency under 150 ms"

After you click 'Run' on this test page you will get results similar to that shown below. You can see that this user is just above these requirements!

You can read more about the Diagnostic Page in Dynamics CRM. You can also monitor the client side using the techniques I describe in my series on Fiddler:

If you take these simple steps to proactively monitor your Dynamics CRM solution then you are much less likely to have a problem that goes un-noticed until you get 'that call'!

@ScottDurow

Pingbacks and trackbacks (6)+

Comments are closed