There I create a BT and filter by application.
My objetive was get all failed transaction for my Application.
Then I compare the results in 2 dashboards.
1 - chart have my custom BT filter By app XXX and select the option "Failed transaction percentage" .
2 - chart have a measure "Failed Transaction" and I filter by dashlet properties and filter by appliction XX.
But the results are different why?
I think this is ok because you are defining Source Measure of a particular Web Request but in 3rd screenshot you are keeping it blank. So in this case the Failed Transaction dashboard data will differ as it will show all error detected in purepaths as for BT dashboard it will only detect the errors when filtering the Web request.
Have you tried to look at associated Purepaths? Your BT metric will count the PP that failed (red icon) while the other one is the PP for your application that had an error (oramge or red icon, not necessarily failed)
Can you post the actual system profile application definition which is used for the mapping of that application:
If you're using something like auto detected ones it may be a different criteria being used to define the application and filtering the application in the BT.
The first Chart have an "Application filter" this application is an auto-detect application.
The second chart have a filter "By uri" the uri is "contains EscritoriosAtyrWeb" the uri is the same that the auto-detect application
If you can I think I would open a case with support, there probably is a logical explanation for this but I can't think of anything that can be tested and they may be able to provide a more detailed explanation (which I would also be interested in).