We have an application(say XYZ) for which we're trying to enable User experience monitoring and facing below issues.
These requests for XYZ application arise from browser à Apache server à IIS Server à WAS JVM. Both Apache and IIS servers are under same agent mapping/group on which UEM sensors are enabled.
We have another application(say ABC) which is hosted on same infrastructure(same web/app servers, same agent mapping/group but different URI) and UEM works just fine for ABC application.
In UEM correlation overview dashboard, on “Dropped/Skipped User Actions” dashlet, we see a spike for ‘UEM - Dropped Actions’(Actions dropped during UEM correlation due to parsing errors or invalid beacon data). Not sure what is wrong.
Since these apps are hosted on same infrastructure, Firewall shouldn’t be blocking Dynatrace/UEM headers.
Wondering what might be causing this issue how to go about narrowing down on the issue.
Are both applications configured to send the beacon to the same path?
Have you accessed the application and watched the network traffic in dev tools or fiddler to see how it behaves?
Are the load balancers for each application configured the same way?
Have you tried capturing the dynatrace headers for each app to verify they are making it through?
After verifying these things, I would open a ticket with support if no clues are forthcoming.
Monitor request path for both applications is different(/XYZ/ vs /ABC/).
Captured fiddler trace while the transactions were being accessed. I can see dynaTraceMonitor signal, dtCookie, other dynaTrace headers etc. on header portion of application URL; See below Fiddler snapshot of response headers on a sample URL of XYZ app;
As part of troubleshooting UEM for XYZ application, I could not find dtCookie value seen on response headers section from Fiddler when searched in Visits dashlet(I was assuming that it would show up as Session ID under visits dashlet)?
However, I can find a matching session in Fiddler when I serach in fiddler using session cookie value under UEM Uplink detilas of purepaths associated with these transactions. This confirms that server side purepaths are being captured. But I can not drill down to visit/user action etc. from these purepaths.
What I can suggest is not keeping two different kind of webserver in single Agent Group. I don’t think it is issue here but it’s not good practice of mixing technologies. But it may be as well kind of issue for those dropped user actions data. Try to create separate agent groups for IIS and Apache. How did you create agent groups for all your agents? Did you use wizard? I’m concerned about having all needed sensor packs placed.