RMFODR - z/OS Service Class Analysis by SYSPLEX, CPCID, and SYSID
The RMFODR inquiry creates charts that display resource usage, performance indicators, summary sampled state percentages, and detailed execution delay state information for each Workload Manager (WLM) Service Class Period (SCP).
mics142
3
2
The inquiry produces vertical bar charts. The SYSID level shows the total CP, zIIP, and zAAP times stacked for each hour of the day. Charts are created at three levels of granularity--z/OS LPAR (SYSID), Central Processing Complex (CPC), and SYSPLEX levels:
- Service Class by SYSPLEX
- Service Class by SYSPLEX CPCID
- Service Class by SYSPLEX CPCID SYSID
At each level of granularity, the following four chart sets are produced, with one chart for each SCP:
1. CPU and PI
2. Sampling Summary
3. Exec Delay States Group 1
4. Exec Delay States Group 2
The title area for all RMFODR inquiry charts shows the SYSPLEX name, Service Class Name, Relative Importance, Goal Type, and Service Class period charted. As delivered, each RMFODR inquiry chart shows the hourly activity for a day. For each SYSPLEX, the Service Class Periods are presented in Relative Importance order, with the most important Service Classes presented first.
With the first chart, CPU and PI, the Performance Index (PI) and the CPU time used by the Service Class Period are shown are shown. The CPU times used for each processor core type (CP, zIIP, and zAAP) are stacked and are scaled using the left Y-axis. The left y-axis is delivered with fixed values over all charts, showing 0 to 5 hours of CPU time. The vertical bars show the total CPU time used by each SCP for the three core types. By using a fixed axis, you can easily identify Service Classes in the chart thumbnail section that used significant amounts of CPU time. If the aggregate amount of CPU time used by Service Classes at your site exceeds 5 hours over one hour of clock time, you will want to adjust the left-Y axis max value higher. Similarly, if the Service Class using the most CPU time in your SYSPLEX uses significantly less than 5 hours of CPU time per clock hour, you will want to adjust the left-Y axis max value lower. Two lines are plotted against the right Y-axis scale - the Performance Index (PI) for the SCP, and a Performance Index "goal met" reference line, set to 1. To prevent scaling issues, PI values greater than 5 are set to 5. PI values between zero (0) and one (1) mean that the Service Class exceeded the specified performance goal. A value equal to 1 means that the goal was met exactly. Values greater than 1 means that the specified performance goal was not met. The PI cannot be computed for certain goal types such as DISCRETIONARY and SYSTEM SPECIFIED.
The second chart, Sampling Summary, displays plays the percentage of times workloads for the SCP were found to be in one of three states: Using, Wait-Unknown, and Execution Delays. The percentage of each state is stacked using vertical bars and are scaled using the left Y-axis fixed to show values from 0 to 100 percent. The Using state represents the percentage of samples where the workloads were using the CPU, and if I/O Priority is active, performing I/O. The Wait-Unknown state represents the percentage of workload delay states that WLM does not manage. These delays include time waiting on an I/O, DB2 latches, tape mount, enqueues, and others. The Execution Delay state represents the percentage of samples where the SCP workloads experienced execution delays for resources that WLM does manage. A fourth state (not shown), if added to the chart would bring the stacked bars up to the 100% line, is the percentage of samples where workloads in the SCP were idle.
The third and fourth charts, Exec Delay States Group 1 and Exec Delay States Group 2, show the percentage of samples belonging to the various categories of execution delays. There are too many categories to effectively display using a single chart, so two chart views are provided, Group 1 and Group 2. The sample percentages of each delay category are stacked using vertical bars against the left Y axis. The axis scale is dynamically adjusted based on the percentages encountered in the data. As with the other chart views, two lines are plotted against the right Y-axis scale -- the performance index (PI) for the Service Class period, and a performance index "goal met" reference line, set to 1.
The first chart sample shown below is an example of the CPU and PI chart at the Service Class by SYSPLEX CPCID level. Shown is a one day hourly chart for Period 1 of Service Class STCDB2. This Service Class executed in SYSPLEX PLEXC1, and the information displayed occurred on z/OS LPARS executing on an IBM 3906-608 CPC. In this chart example, you can see that workloads in the STCTEST Service Class used both CP and zIIP processor cores, as shown by the blue and red stacked bars. Service Class STCDB2 has an execution velocity goal type. The actual performance index value, shown with a light green line, is generally below the reference value of 1, meaning that the actual performance of workloads running in this Service Class generally exceeded the specified goal. But at 10 AM and 2 PM, the PI was value over 2, meaning the goal was missed.
The second chart sample shown below is an example of the Exec Delay States Group 1 chart at the Service Class by SYSPLEX CPCID level. This chart sample is generated from the same data used for the first chart sample. The chart shows the first set of execution delay category percentages for Period 1 of the Service Class STCDB2. You can see that the reason the SCP goal was missed for hours 10 AM and 2 PM was primarily due to an increase in DASD I/O delays for the STCDB2 workloads.


Uses
All workloads running on z/OS are assigned to a Service Class, and each Service Class is defined with a level of importance. Most Service Classes are also assigned a goal type (for example, Execution Velocity, Response Time Percentile, and so on) and a performance goal to meet. The z/OS Workload Manager (WLM) manages physical resources (for example, CPU) and uses workload dispatching strategies to meet Service Class goals. Service classes with a higher relative importance value (0 high <--> 5 low) are given preferential treatment.
All charts generated by the RMFODR inquiry are organized by SYSPLEX, and charts are produced for each Service Class Period. The CPU and PI charts show the CPU utilization and Performance Index (PI) for each HOUR. The PI indicates how well the workload performed against the designated goal.
This CPU and PI chart provides the information required to understand the performance and resource consumption of z/OS workloads. Using this chart, you can monitor the performance of each Service Class Period defined in your WLM Service Policies and Definitions. If a Service Class is routinely exceeding the assigned goal (for example, the PI is consistently 0.8 or lower), or missing the assigned goal (for example, the PI is consistently higher than 1), you can adjust the goal or modify other Service Classes to free up additional resources for the troubled Service Class.
The second chart, Sampling Summary, allows you to quickly determine whether poorly performing SCPs are being delayed by WLM managed resources, or for other reasons. If you see increased WLM execution delays at times when the SCP goals are missed, you can examine charts 3 and 4 to determine the specific types of execution delays that impacted performance. The resources that are represented by these delay categories are managed by WLM. The delays that are experienced by Service Classes can be alleviated by adjusting Service Policy Service Class goals or by adding additional capacity of the resource type indicated.
By design, SYSPLEXES can span Central Processing Complexes. If your SYSPLEXES do not span CPCs, then the first chart levels are identical, with the exception that the second chart level, Service Class by SYSPLEX CPCID, identifies the CPC in each chart title. If your SYSPLEXES do span CPCs, then the Service Class by SYSPLEX CPCID level displays multiple charts for a Service Class that ran on multiple CPCs.
Charts - Service Class by SYSPLEX
VIEW 1 of 4: WLMCLASS by SYSPLEX - CPU and PI

Note:
RCP = Report Class PeriodLeft Y-axis
- SECCPUTM
- CPU Time on CP
- SECSUPTM
- CPU Time on zIIP (Normalized)
- SECZAPTM
- CPU Time on zAAP (Normalized)
Right Y-axis
- SECPINDX
- Performance Index
- Performance Index Reference - value is set at 1
X-axis
- HHMM
- Hour: Minute
VIEW 2 of 4: WLMCLASS by SYSPLEX - Sampling Summary

Left Y-axis
- SEDPCUSE
- Pct Total WLM Using Samples (derived in inquiry code)
- SEDPCDUN
- Pct Wait - Unknown
- SEDPCSAM
- Pct Total Execution Delay Samples (derived in inquiry code)
Right Y-axis
- SECPINDX
- Performance Index
- Performance Index Reference - value set to 1
X-axis
- HHMM
- Hour: Minute
VIEW 3 of 4: WLMCLASS by SYSPLEX - Exec Delay States Group 1

Left Y-axis
- SEDPCDCP
- Pct Wait - CPU on CP
- SEDPCSUD
- Pct Wait - CPU on zIIP
- SEDPCZPD
- Pct Wait - CPU on zAAP
- SEDPCAMD
- Pct Wait - Crypto AM Processor
- SEDPCAPD
- Pct Wait - Crypto Assist Processor
- SEDPCCRD
- Pct Wait - Crypto Processors
- SEDPCCOD
- Pct Wait - Resource Contention
- SEDPCDCC
- Pct Wait - CPU Capping
- SEDPCDQU
- Pct Wait - Quiesce
- SEDPCDMP
- Pct Wait - MPL Delay
- SEDPCSMP
- Pct Wait - Server MPL
- SEDPCDRG
- Pct Wait - Resource Group Capping
- SEDPCDSD
- Pct Wait - DASD I/O Delay
- SEDPCSQU
- Pct Wait - Server Queue Delay
Right Y-axis
- SECPINDX
- Performance Index
- Performance Index Reference - value set to 1
X-axis
- HHMM
- Hour: Minute
VIEW 4 of 4: WLMCLASS by SYSPLEX - Exec Delay States Group 2

Left Y-axis
- SEDPCDES
- Pct Wait - ESO Hiperspc Pg-In from Aux
- SEDPCDHS
- Pct Wait - Std Hiperspc Pg-In from Aux
- SEDPCDPC
- Pct Wait - CSA Page-In from Aux
- SEDPCDPP
- Pct Wait - Pvt Page-In from Aux
- SEDPCDSI
- Pct Wait - Swap-In
- SEDPCDVI
- Pct Wait - VIO Page-In from Aux
- SEDPCDXM
- Pct Wait - Cross Mem Page-In from Aux
- SEDPCSHP
- Pct Wait - Server Hiperspace Paging
- SEDPCSPP
- Pct Wait - Server Private Area Paging
- SEDPCSSI
- Pct Wait - Server Swap-In
- SEDPCSVP
- Pct Wait - Server Space VIO Paging
Right Y-axis
- SECPINDX
- Performance Index
- Performance Index Reference - value set to 1
X-axis
- HHMM
- Hour: Minute
Service Class by SYSPLEX CPCID
Data extract Service Class by SYSPLEX CPCID provides the same chart views, with charts displaying the same Y and X-axis charted data elements, as those described for data extract Service Class by SYSPLEX.
The only difference is that additional charts are generated if the SYSPLEX spans multiple CPCs. Where the chart granularity for the SYSPLEX level data extract followed this pattern:

The chart granularity for the Service Class by SYSPLEX CPCID chart follows this pattern:

If a SYSPLEX spans two or more CPCs, then the SYSPLEX CPCID level will generate more unique charts that the SYSPLEX level.
If a SYSPLEX does not span CPCs, however, the number of unique charts generated at the SYSPLEX level and SYSPLEX CPCID level will be identical.
Service Class by SYSPLEX CPCID SYSID
The Service Class by SYSPLEX CPCID SYSID provides the same sets of charts, with each chart displaying the same Y and X-axis data elements as those described for data extract Service Class by SYSPLEX.
The only difference is that additional charts are generated if the SYSPLEX spans multiple CPCs, or multiple z/OS LPARs.
Where the chart granularity at the SYSPLEX level follows this pattern:

The chart granularity at the Service Class by SYSPLEX CPCID SYSID level follows this pattern:

At this level, you can examine the performance of each Service Class Period on each z/OS LPAR.
Inquiry Defaults
DATA SOURCE (file/timespan):
- WLMSEC at the DAYS timespan
- WLMSED at the DAYS timespan
- Filters: Empty Filter tasks provided as placeholder for user WLMSEC and WLMSED file filtering to exclude/select specific SYSPLEXs, SYSIDs, ZONEs, and so on.
- Derivations:
CPCID
%CPCID ;
DATE
DATEPART(ENDTS);
HHMM
PUT(HOUR(ENDTS),Z2.) || ':' || PUT(MINUTE(ENDTS),Z2.);
DATEHOUR
PUT(DATEPART(ENDTS),YYMMDDD10.) || ':' || PUT(HOUR,Z2.) ;
DATEHHMM
PUT(DATEPART(ENDTS),YYMMDDP10.) || '-' || HHMM ;
SECFILE
1 ; Used after MERGE of WLMSEC and WLMSED files to limit charting to observations that contained WLMSEC file data.
SEDPCSAM
100 * SEDTOSAM / SEDTOSST;
SEDPCUSE
100 * SEDTOUSE / SEDTOSST;
Run-Time Execution Parameter Overrides:
- UNIT
- Select one or moreMICS databaseunits by entering the data base IDs. Enter multiple IDs with a blank following each ID value.
- CYCLE RANGE
- Enter range ofMICSonline file cycles or enter ARCHIVE for archive files.
- SELECT YESTERDAY
- Specify whether to report only on observations with a time stamp from yesterday.
- Available values:
- Y
- - Yes
- N
- - No (default)
- SYSPLEX
- Enter SYSPLEX name or leave blank to chart all SYSPLEX's.
- LEFT-Y AXIS SCALE
- Specify the maximum Left-Y axis chart value for the Service Class CPU and PI charts. The default is 5 hours. If you have report classes that consistently exceed 5 hours of CPU time in an hour, you can increase the maximum scale, or lower the scale to a value that would be more representative of your workloads. Specifying a value of zero (0) removes the maximum scaling and lets SAS adjust the scale based on the data charted.
- 5 - Maximum Left-Y axis default
- 0 - SAS automatic scaling
- Recommended Execution Time parameter setting for daily executions:
- Set Cycle range to 01-02
- Set Select Yesterday to Y