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

Mobile APP UEM

kyle_kowalski
Dynatrace Pro
Dynatrace Pro
Does anyone have any experience with instrumenting a mobile app with UEM data. We are having some issues having it contacting our dynaTraceMonitor. We see data being recorded in the log files but don't see the visit in Dynatrace AM.
Kyle
17 REPLIES 17

KlausEnzenhofer
Dynatrace Pro
Dynatrace Pro

Hi Kyle,

Which dt version are you on? With the logs you mean the mobile app log? Have you checked the correlation log on the dt server for any suspicious warnings that data gets dropped?

Klaus

kyle_kowalski
Dynatrace Pro
Dynatrace Pro

Thanks for your time Klaus, we are using dynaTrace version 6.1.0.7880, yes it was the mobile application logs where we were seeing performance data and user actions being recorded, I haven't gotten a chance to view the server logs yet which one should I be looking at for this?

Thanks,

Kyle

 

edit: Updated version information to current build

The server correlation log file. In case there is an issue and the data gets dropped you will see a message there.

Klaus, which file is this specifically? or at least where is it located? I do not see it in the system information dashlet in the client and I am trying to troubleshoot some mobile communication issues. We have a web server backend for the mobile app that is working jsut fine but we suddenly stopped receiving UEM from the mobile application itself.

carlos
Inactive

Hi, the correlation log file will be at System Information, just look for the DT server logs. One of them is the server correlation.

Regarding UEM, if you browse from a mobile browser your web server injecting the J script, can you reach whatever/dynaTraceMonitor? (or look at the source coude there for the dtagent injected)?

Just to make sure you have mobile to web server + injection working first!

Bests,
Carlos.

Thanks, some user error going on over here as I had searched for it there but must have typoed.

Navigating to the URL gives me a 401 unauthorized access warning for invalid credentials.

EDIT: Actually turns out if you use find before the System Information full loads, it will not load anymore so it will appear as though your search shows nothing even though it finds something.

Great, just a permission issue, easy to correct.

this is the same result I get when I try to access our production URL as well but production works despite the "unauthorized" error.

carlos
Inactive

Ok, but if you look into the source code in the browser, do you find the injected script from Dynatrace? Is it there?

No even in production the monitor URL is not taking me to the page that I can see that. It only takes me to the cannot be authenticated page, which has nothing showing the script injected.

We have a download page for the application itself that is also within the same APP pool and that has the Dynatrace URL monitor in the source code in both dev and production.

I do not know if I'm understanding correctly, but in any case, any html loaded from the application defined under UEM setting, should show you the Java script.

I'w add POST/GET permissions to i.e. Everyone on the dynaTraceMonitor path of your web app/server. In order to avoid the 401 codes, which may be potentially affect measures, information delivery, etc...in my humble opinion.

Also, at least we should make sure the "sliding bar" on UEM setting is in the position to show you End User Experience etc...not only server side. Did you check that? Or if you can post that screenshot for the given application and environment.

jared_kemp
Organizer

The sliding bar is set for the UEM on all the way to the right. We are injecting the code as I see the mobile browsers downloading the application for that page and they are part of the same agent group, I am just not getting any data from the app itself. The mobile browser and top desktop browser are both for that agent group, UEM appears to be working for them, just not involving any data from the mobile application itself.

carlos
Inactive

Ok, so you don't see any "visits" basically? And not even user actions I assume?

How about if you define the Mobile Application for UEM? I can see it is listed in the transaction flow, shown as Default Application. Have you defined the app?

Yes, no visits or user actions for the mobile app, I only see user actions for the IIS portion of the application which is just app download and a help feature.The application name is all defined via the mobile ADK itself, but since no data is coming from the ADK no app appears. Now the default application for the mobile browser is because it goes through IIS and sadly you cannot assign an application defined by an ADK to also be defined manually.

carlos
Inactive

Ok, now, the mobile app is it native Android or iOS built? Or do you (or the customer) use any framework to develop/build the mobile app)?

Based on your previous comments, it was working as expected and suddenly there were no visits at all? Is this correct?

Here is a screenshot from a mobile app, we used auto-instrumentation, even though the mobile app was built with Cordova Framework.

If you see the requests from mobile users to the web server, you should be able to define that application, not sure if this solves the issue, but it is a healthy practice though. In my example application xxxx QAS xxx was defined.

it is an iOS app that was built. I am still waiting on the devs tog et back to me with log info so I am trying to sure up everything on my end. Yes, teh applciation was working then suddenly stop working over the weekend with no apparent changes from the Devs and I made none on the dynatrace side.

I do know I can set an application from the web server but I cannot assign the same application as the mobile app so I had not made anything for the dev environment yet.

carlos
Inactive

I understand, it will be great to take a look at the devs logs. If you can reach the URL defined to inject the Java script, from the mobiles, everything is almost ready.

Or you could also debug the mobile app on a device, i.e. with log cat, to make sure you can see the requests and info targeting that URL.

Did it stop working gradually (less and less visits during the weekend or at once? Trying to think if there was a new roll out involved or an infrastructure issue/change...