SUMMARY:
This article provides a detailed description of how results are calculated by the Enterprise Manager Metric ''RP Average Execution Duration per Job (hh:mm:ss)'. This article provides a way to verify results from the metric are accurate.
DEFINITION OF METRIC:
([RP Exec Duration (hh:mm:ss)]/[RP Number of Jobs])
The metric is calculated based on the average time for every job that was run in the desired timeframe. Since the metric gets an average, it is necessary to consider all the executions of that specific report in the given time period. For example, if a report took 4 hours to execute once and then was executed two more times with execution times of 1 minute and 10 minutes respectively, it would be inaccurate to expect the average will be 4 hours. Instead it will be approximately 1.23 hours.
This can be tested with any out of the box Enterprise Manager report that uses this metric but for the following example, the Report No. 6 "Top 10 Longest Reports" was used.
2) Job Indicator: Move Cancelled Jobs, Error Jobs, and Jobs that have timed out to the left
3) Click Next
4) Click Finish
The report will run.
5) Select a report to use to verify the metric is correct
6) Use the 'Search for Objects' tool to find the report and its properties
7) In DB Query Tool perform the following statement SELECT * IS_REP WHERE IS_REP_NAME = '<Report Name>'
8) Get the Job ID
9) In DB Query Tool perform the following statement SELECT * FROM IS_REP_FACT WHERE IS_REP_JOB_ID = '<Job ID>'
10) Add the values in IS_REP_EXEC_TM_MS that fall within the last month (Note the values are recorded in milliseconds)
12) Divide the aggregate by the number of columns corresponding to the last month
This test verifies how the metric RP Average Execution Duration per Job (hh:mm:ss) calculates.
