How do I tie exceptions into our Apdex rating?

[kjordan]

In our environment if an exception happens we capture it and redirect the user to a landing error page shared by the company. That redirect is happening in our C# code where the exception has been caught. The error page is in an Enterprise application. We have other applications in Dynatrace that are broken out by system, each with its own Apdex.

What I'm looking to do is have those exceptions actually impact the Apdex rating for the application it's coming from, not the Enterprise application it's landing in. I know that I could throw a JS error of some kind on the error page and it would impact Apdex, but that just impacts the Apdex for the Enterprise application rather than the application the request is coming from. I'm already in the C# code at the time it's redirecting to the error page so I can't fire a JS error on the page it's routing from.

I tried looking at the Dynatrace API but don't see anything in there that would allow me to report an error like this to a given application. The JS API has a reportError function but that doesn't help me when I'm already in the C# code by the time I know an exception has happened. I can't call reportError in JS from the page the error occurred on.

Has anyone else run into this? Does anyone have any suggestions for what I could do to be able to feed the apdex the information from these? As it is it actually would improve the apdex if we started having serious system issues because requests would stop short and they'd redirect to a very quickly loading error page.

[Julius_Loman · Accepted solution]

Web Applications in Dynatrace are determined by the URL. Mobile application / OpenKit applications similarly by using an application ID. In Dynatrace application term is the "user facing" side of an application - user perception. Application entity in Dynatrace measures the end-user experience only.

Basically I understood you have an error status in a backend application, but it's feeding data for a Web Application. Correct? In this case, the user is accessing the Web Application and the error should be thrown there using the Dynatrace RUM JavaScript API - he/she accesses the web application.

On the server-side - the exceptions/errors can be handled on the service settings. If you know your request has an erroneous state, you should mark the request as failed. You can tune the rules on how requests are marked as failed. For example when a particular exception occurs in the processing of the request, request is marked as failed. However, this won't have any impact on apdex.

Hope it helps.