[ https://issues.jboss.org/browse/JBIDE-20167?page=com.atlassian.jira.plugi... ] Alexey Kazakov edited comment on JBIDE-20167 at 8/6/15 10:15 AM: ----------------------------------------------------------------- Reopened to clarify. I see that you use a value of events to track failures=0 and success=1. But when you are *counting* events using UsageReporter.getInstance().countEvent() we send the only one event during 24 hours. This event which we actually send will have value = sum of all values of events you have counted. So if you count five events (two success, three failures, sum of values = 1+1+0+0+0) then we will send one event with value = 2. It's impossible to track success/failures using this approach. You should use values of counted events only for tracking the number of real user events. Not for something else. You have to *track* events instead of counting them (it means you will send an event immediately with value=0 or 1) so you will have the number of events=number of wizard launches and value of these events will show the % of success/failure. was (Author: akazakov): Reopened to clarify. I see that you use a value of events to track failures=0 and success=1. But when you are *counting* events using UsageReporter.getInstance().countEvent() we send the only one event during 24 hours. This event which we actually send will have value = sum of all values of events you have counted. So if you count five events (two success, free failures, sum of values = 1+1+0+0+0) then we will send one event with value = 2. It's impossible to track success/failures using this approach. You should use values of counted events only for tracking the number of real user events. Not for something else. You have to *track* events instead of counting them (it means you will send an event immediately with value=0 or 1) so you will have the number of events=number of wizard launches and value of these events will show the % of success/failure. -- This message was sent by Atlassian JIRA (v6.3.15#6346)