RMF Appendix 2: Tips for Doing Ad Hoc Analysis | |
Although the examples shown here are still valid, you may prefer to use a new feature, "Graph Data," which is available in the IT Service Vision client PC interface and has equivalent functionality
In this example you will use the IT Service Vision server's drill chart feature to investigate CPU utilization.
PDB Admin -> Manage PDBs/Switch PDB -> select the demonstration PDB -> ItemActions -> Activate -> Read Access -> OK -> File -> End
Reports -> Adhoc Analysis -> Drill Chart -> DETAIL -> XTY70
Select Goback.
View -> Variable -> ID value
and scrolling until you see the SHIFT variable. Select the SHIFT variable. The bars will now split into different colors, where each color represents a different shift, as described in the legend.
Select
View -> Variable -> Y axis
Now you could select, say, IORATE or TSOAVG. Use this facility to explore your data.
In this example you will use the IT Service Vision server's graph data feature to analyze the CPU consumption of different jobs and users.
PDB Admin -> Manage PDBs/Switch PDB -> select the demonstration PDB -> ItemActions -> Activate -> Read Access -> OK -> File -> End
Reports -> Adhoc Analysis -> Graph Data -> DETAIL -> XSTEPS
Variable -> RAFGRUP -> Operator -> IN -> Value(s)
IT Service Vision lists all the different values that RAFGRUP has within the XSTEPS table. Select the following values, scrolling when necessary:
GER, ITA, SCH, SNL, SSF, SUK
Then select OK and OK.
The scatter plot will then re-display and show observations for only these RACF groups.
Clicking on a bar gives information about the statistical make-up of that bar.
Select the "Sigma" button and change the statistic to Sum. Clicking on a bar will now show the new statistic.
In the sample data there is a one-to-one mapping of RAFUSER to RAFGRUP, so the ID variable shows no great advantage in this example. This feature would best be shown on some of your own data, where each bar could represent an account code and an ID variable could be the userid. Then each bar would be broken into subgroups (userids). Or you could specify SHIFT as an ID variable.
Now that you understand this facility, you can experiment with your own data.