We have questions about the differences in the reports. The aggregator report shows actual hourly averages. We are not sure how iDash calculates the averages. The aggregator in our environment shows average latency as 1 to 2 seconds, but iDash sometimes shows minutes.
Can you please tell us how latencies are calculated?
We get the que_status_stamp and event_time_gmt values from ujo_proc_event followed by subtraction ( que_status_stamp – event_time_gmt) after converting the que_status_stamp to Integer.
Then, int value = (Math.abs(processTime - postTime)) % 3600 where processTime is the converted que_status_stamp and postTime is event_time_gmt and 3600 is 60 minutes to offset timezone value where TZs have a 30 minute offset like IST (+5:30).
After this if there is a timezone offset between iDash server and AutoSys Instance TZ configured in iDash Admin Tool then the calculation is:
value = processTime - postTime – tzOffset.
We avoid the negative latency so we will pick values where “value > 0 from the above calculation”. This will be in a SQL resultSet.next() loop so for each iteration we will add the “minuteInTime” value and the corresponding values where computed “value is > 0” as above.