Showing results for 
Show  only  | Search instead for 
Did you mean: 

This product reached the end of support date on March 31, 2021.

How to monitor failing web service calls?



Initially I used a soap analyzer to monitor a certain web service but at after a while we had a service failure and I realized that if the request and response do not match the extraction of operation name, user name, etc. does not work. As a consequence I could not give any details about the failing calls.

Then I changed to the xml analyzer where I can define separate operations that match the way failures may look but then again Gomez does not regard those failing operations as failures anymore. I do handle this with custom reports but sadly all the standard reports will be wrong because of this.

How can I get both my details on the failing calls and keep using standard reports?

Seems to me it might be handy to have a way to state that even if a certain operation matches a request/response pair it is to be considered as a failure nonetheless.

Maybe I have missed something obvious?

Kind regards





Are you referring to the Compuware SaaS Synthetic Web Monitoring or the Enterprise Synthetic Monitoring?

Based on the question, it appears to be more directly related to the Enterprise product, but I want to conform that.


Thanks for replying.

I'm not sure how to answer your question. The tool was sold to my company as 'Gomez' and the 'About page' says 'Central Analysis Server, version:'

The choice of analyzer and configuration of it is done with 'RUM Console.'

If I pick the soap analyzer web requests and responses must follow rather strict syntax or be considered a discarded call. If I pick the xml analyzer instead I am able to map the various variants in replies corresponding to different error scenarios. But doing this also makes Gomez consider the calls succesfull.

So, how can I get the various failure responses (for statistic reasons) without missing the fact they are indeed failures?


When your tansactions fail, do they all present the same error or failure information?

In this case there are a few different variations of what errors look like but whether I can match them with patterns or not is besides the point I think because once they are matched they are considered successful calls I believe.


Hi Michael,

What product version are you running?

Our 12.0 version supports the reporting of standard SOAP fault errors and recently there was an improvement in one of the latest 12.0 builds to report application fault errors as well. From what I understood from your issue description, I believe upgrading to the latest 12.0 build would resolve this issue.

If you are already on 12.0, please contact our support team to open a case and reference this thread and we will be glad to provide you with the latest build with the SOAP fault improvements. If you are not on 12.0 yet, I suggest upgrading first and then contact us to obtain the latest 12.0 build. If you need assistance in upgrading to 12.0, please check out our migration center (



Well, the about page says CAS but I am glad to learn 12.0 includes improvements in this area.

We are in the process of planning upgrading to 12.0 so maybe I should just wait for this to happens.

Thanks all for replying!



Good to know that you are planning an upgrade to 12.0. Again, from what you have described, I believe that the latest 12.0 build we currently have would resolve your issue. After you upgrade to 12.0, please contact our support team to obtain the latest 12.0 build which contains the SOAP fault improvements and reference this post.

Another thing that you need to do after the upgrade to the latest 12.0 build is to switch back the software service to use the SOAP analyzer instead of XML in order for the operations to report correctly.




Hi, some SOAP errors which are "business related" give the 500 Internal server error code. This is however correct according to the specification of a WS.

Will the SOAP analysis module take this into consideration? (PS running on DCRUM 12.3.4)


You can exclude particular type of errors from calculating Availability metric:

Is it what you're looking for?