CRITICAL |
This indication reports on the specification in the service policy, not whether SRM is using
or ignoring it. This information can be found in Monitor II (ARD report) and in Monitor III (for
example, DELAY report).
- CPU
- CPU critical
- STORAGE
- Storage critical
- CPU + STORAGE
- Both CPU and storage critical
- NONE
- Neither storage nor CPU critical
|
HONOR PRIORITY=NO |
This indication is reported for service classes that do
not allow to overflow to regular CPs when there is insufficient zIIP or zAAP capacity. |
I/O PRIORITY GROUP=HIGH |
This indication is reported for service classes assigned to I/O priority group HIGH in the
active service policy. |
TRANSACTIONS |
Number of transactions.
- AVG
- The average number of active transactions during the interval including independent enclaves. On
a period level, this is the transaction active time for the single period, divided by the RMF
interval time. Otherwise, it is the sum of the transaction active time for all summarized periods,
divided by the RMF interval time.
To get the average number of transactions related to active
address spaces, subtract AVG ENC from AVG.
- MPL
- The average number of transactions resident in central storage during the interval. On a period
level, it is the transaction residency time for a single period, divided by the RMF interval time.
Otherwise, it is the sum of the transaction residency time for all summarized periods, divided by
the RMF interval time.
To get the average number of transactions related to resident address
spaces, subtract AVG ENC from MPL.
- ENDED
- The number of transactions that ended during the interval. On a period level, this is the number
of transactions that ended during that period. Otherwise, it is the total number of transactions
that ended for all the summarized periods.
- END/S
- The number of transactions that ended per second.
- #SWAPS
- The total number of swaps. On a period level it is the number of swaps occurred during the
single period. Otherwise, it is the sum of the number of SWAPS that occurred during all summarized
periods.
- EXCTD
- Count of times a subsystem work manager reported that an execution phase has completed. A single
transaction could have zero or more execution phases.
|
TRANS-TIME |
Transaction time in HHH.MM.SS.FFFFFF
units. All times are reported in the period the transaction ended. The time a job was delayed due
to TYPRUN=HOLD or TYPRUN=JCLHOLD is NOT included in any of the transaction times.
- ACTUAL
- The actual amount of time required to complete the work submitted under the service class. This
is the total response time including EXECUTION, QUEUED, R/S AFFIN, and INELIGIBLE.
- EXECUTION
- The average execution time of ended transactions.
- QUEUED
- Average time a job was delayed for reasons other than the ones mentioned below. This field
therefore basically includes the time a job was delayed for initiation.
For TSO users, this can
be a portion of LOGON processing. For APPC this is the time the transaction spent on an APPC queue.
- R/S AFFIN
- Average time the job was delayed due to resource or system affinity scheduling delay. This means
that resource(s) required for the job to run were not available at some point while the job was
queued to JES2.
- INELIGIBLE
- Average time the job was delayed due to operational delays or JES scheduling delays, examples
are:
- Job held by operator
- Job class or job queue held
- Duplicate jobname serialization
- Job class execution limits
- CONVERSION
- Average time the job was delayed due to JCL conversion.
Jobs held during conversion (due to
affinity, HSM recall, or enqueue contention) contribute only to conversion time, not to ineligible
or R/S affinity times.
CONVERSION time is not included in the total response time (ACTUAL).
- STD DEV
- Standard deviation of ACTUAL.
Standard deviation is a measure of variability of the data in
the sample. The higher the standard deviation, the more spread out it looks on a graph.
|
Transaction Application Time % (TRANS-APPL%)
When transaction processor usage is reported to WLM through IWM4RPT or IWM4MNTF services, the
consumed service units are accounted to the transaction service or report classes, and deducted from
the region's service and report classes. If the number of transactions is very small and a single
transaction reports high processor times, it can occur that processor times become negative. In such
a case RMF displays asterisk (*).
|
TOTAL |
Total percentage of the processor time used by transactions running on the different
processor types.
- CP
- Total percentage of general purpose processor time used by transactions.
- IIPCP/AAPCP
- Total percentage of general purpose processor time used by transactions eligible to run on
specialty processors.
- IIP/AAP
- Total percentage of specialty processor time used by transactions.
|
MOBILE CATEGORYA CATEGORYB
|
Percentage of the processor time used by transactions classified with
reporting attribute MOBILE / CATEGORYA / CATEGORYB running on the different
processor types. MOBILE / CATEGORYA / CATEGORYB is a subset of TOTAL.
- CP
- Percentage of general purpose processor time used by transactions classified with reporting
attribute MOBILE / CATEGORYA / CATEGORYB.
- IIPCP/AAPCP
- Percentage of general purpose processor time used by transactions classified with reporting
attribute MOBILE / CATEGORYA / CATEGORYB, eligible to run on specialty processors.
- IIP/AAP
- Percentage of specialty processor time used by transactions classified with reporting attribute
MOBILE / CATEGORYA / CATEGORYB.
|
ENCLAVES |
Average number of enclaves.
- AVG ENC
- The average number of independent enclaves during the interval. From a sysplex scope, this is
the sum of active time for enclaves that originated on the respective system either for the single
period or for all summarized periods divided by the RMF interval time.
- REM ENC
- The average number of foreign enclaves during the interval. From a sysplex scope, this is the
sum of active time for enclaves that originated on a remote system in the sysplex, but are executing
on the respective system either for the single period or for all summarized periods divided by the
RMF interval time.
- MS ENC
- The average number of multi-system enclaves during the interval. From a sysplex scope, this is
the sum of active time for enclaves that originated on the respective system and are executing on
one or more remote systems in the sysplex in parallel either for the single period or for all
summarized periods divided by the RMF interval time.
|
DASD I/O |
Information about DASD I/O activities.
- SSCHRT
- Number of start subchannels SSCH per second in the reported interval.
- RESP
- Average DASD response time (in milliseconds) of the transactions in this group. This is the sum
of the average connect time (CONN), the average disconnect time (DISC), the average wait time
(Q+PEND), and the IOS queue time (IOSQ).
- CONN
- Average DASD connection time of the transactions in this group, as reported by the channel
measurement subsystem.
- DISC
- Average DASD disconnect time of the transactions in this group, as reported by the channel
measurement subsystem.
- Q+PEND
- Average DASD wait time (queue time + pending time) of the transactions in this group. This does
not include IOSQ time, as reported by the channel measurement subsystem.
- IOSQ
- Average time the transactions in this group spent on the IOS queue, based on sampled
delays.
|
SERVICE |
The service is calculated by multiplying the received service units with the appropriate
service definition coefficient.
- IOC
- Total amount of input/output service received.
- CPU
- Total amount of task and preemptible-class SRB processor service received.
- MSO
- Total amount of main storage occupancy service received.
- SRB
- Total amount of non-preemptible SRB service received.
- TOT
- Sum of CPU, SRB, IOC, and MSO service.
- /SEC
- Rate at which service is provided in service units per second.
- ABSRPTN
- Absorption rate at which service is used while transactions are resident in main storage. This
is the total service divided by the transaction residency time.
- TRX SERV
- Rate at which service is used by transactions that are active, but not necessarily in storage.
This is the total service divided by the transaction active time.
|
SERVICE TIME |
This category is made up of the following:
- CPU
- Task and preemptible-class SRB (enclave) time in seconds consumed on general purpose and special
purpose processors.
- SRB
- Service request block time in seconds.
- RCT
- Region control task time in seconds.
- IIT
- I/O interrupt time in seconds.
- HST
- Hiperspace service time in seconds.
- AAP
- zAAP service time in seconds.
- IIP
- zIIP service time in seconds.
Note:
- If special purpose processors are running faster than general purpose processors, AAP and IIP times are not normalized.
- Normalized AAP and IIP times are included in CPU
time.
|
APPL% |
Percentage of the processor time used by transactions running on the different processor
types. The calculation is:
Processor time used
APPL% = -------------------------------------------------------- * 100
Interval length * Multithreading maximum capacity factor
- CP
- Percentage of the processor time used by transactions running on general purpose processors in
the service or report class period. The calculation of the processor time is based on the time
values displayed under field heading SERVICE TIME.
Processor time used = CPU + SRB + RCT + IIT +
HST – (AAPNF * AAP) –
(IIPNF * IIP)
The AAP and IIP times may be
normalized to general purpose processor time from a faster zAAP
or zIIP where AAPNF and IIPNF are the zAAP and zIIP normalization factors. They can be found under field
heading NORM FACTORS on the POLICY page of the Workload Activity report.
- AAPCP
- Percentage of the processor time used by zAAP eligible transactions running on general purpose
processors. This is a subset of APPL% CP.
- IIPCP
- Percentage of the processor time used by zIIP eligible transactions running on general purpose
processors. This is a subset of APPL% CP.
- AAP
- Percentage of the processor time used by transactions executed on zAAPs in the service or report
class period.
- IIP
- Percentage of the processor time used by transactions executed on zIIPs in the service or report
class period.
Notes:
- APPL% shows processor utilization based on uniprocessor capacity. This means that the values can
exceed 100% in systems with more than one processor.
- If the multithreading mode is set to 1, a multithreading maximum capacity factor of 1 is used
for the APPL% calculation.
- The interval length in a sysplex is the common interval length.
- In a sysplex, the values for seconds and CPU time percentages are meaningful only if all
processors have the same speed and the multithreading mode is the same on all systems. You can use
the SYSRPTS WLMGL SYSNAM option to select only a subset of the systems to be included in the
report.
- AAPCP or IIPCP may report values greater than
zero even if no special purpose processors are configured or if they are varied offline, because the
PROJECTCPU option is specified in the active IEAOPT Parmlib member. This information can be used to
understand the benefit of adding special purpose processors to your system.
|
PROMOTED |
CPU time in seconds that transactions in this group were running at a promoted dispatching
priority, separated by the reason for the promotion:
- BLK
- CPU time in seconds consumed while the dispatching priority of work with low importance was
temporarily raised to help blocked workloads
- ENQ
- CPU time in seconds consumed while the dispatching priority was temporarily raised by enqueue
management because the work held a resource that other work needed.
- CRM
- CPU time in seconds consumed while the dispatching priority was temporarily raised by chronic
resource contention management because the work held a resource that other work needed
- LCK
- In HiperDispatch mode, the CPU time in seconds consumed while the dispatching priority was
temporarily raised to shorten the lock hold time of a local suspend lock held by the work unit.
- SUP
- CPU time in seconds consumed while the dispatching priority for a work unit was temporarily
raised by the z/OS supervisor to a higher dispatching priority than assigned by WLM.
|
STORAGE |
Amount of storage frames.
- AVG
- Weighted average number of central and expanded storage frames allocated to active ASIDs. This
value is the sum of the number of central and expanded frames weighted by the transaction residency
time for each active ASID, divided by the total transaction residency time.
Note: Enclave
transaction residency or active time is not included in the calculation of this value.
- TOTAL
- Total number of central and expanded storage frames allocated to resident ASIDs. This value is
the sum of the total number of central and expanded frames weighted by the transaction residency
time for each active ASID, divided by the RMF interval time.
- SHARED
- Total number of shared storage pages allocated to resident ASIDs.
|
PAGE-IN RATES |
- SINGLE
- The average rate at which pages are read into central storage while transactions are resident in
central storage. On a single period level this is the total number of page-ins during the period,
divided by transaction residency time. For all other levels it is the sum of the total number of
page-ins for all periods summarized, divided by the sum of the transaction residency time for all
periods being summarized.
- BLOCK
- Rate of demand page-ins from DASD for blocked pages, expressed in pages per seconds.
- SHARED
- Rate of shared storage page-ins
- HSP
- Rate of standard hiperspace pages read into central storage from auxiliary storage.
Note: Enclave transaction residency time is not included in the calculation of these values because
there is no paging on behalf of enclaves.
|
MOBILE |
Percentage of the processor time used by transactions classified with
reporting attribute MOBILE running on the different processor types. MOBILE is a subset of TOTAL.
- CP
- Percentage of general purpose processor time used by transactions classified with reporting
attribute MOBILE.
- AAP/IIP ON CP
- Percentage of general purpose processor time used by transactions classified with reporting
attribute MOBILE, eligible to run on specialty processors.
- AAP/IIP
- Percentage of specialty processor time used by transactions classified with reporting attribute
MOBILE.
|
Service Classes being Served
|
SERVICE CLASSES BEING SERVED |
This section is only available if address spaces are doing work for transactions that were
classified to another service class. The name of each service class being served by the reported
service class (see name in the separation line) is displayed. |
Work Manager/Resource Manager
|
SUB TYPE |
The name (for example CICS or IMS) represents the subsystem type (4 characters) as used in
the classification rules in the WLM administration application. The subsystem's documentation should
explain the meaning that product attributes to the specific states. |
P |
The phase identified as BTE indicates the representation of the states incurred in the
begin-to-end phase of a transaction EXE indicates the representation of the states incurred
in the execution phase of a transaction. |
RESP TIME (%) |
The transaction response time percentage in either the BEGIN-TO-END phase, or
the EXECUTION phase. |
STATE SAMPLES BREAKDOWN (%) |
Identifies the percentages of samples that a transaction has been detected in the reported
states.
- ACTIVE SUB
- The active subsystem state sample percentage. Active indicates that there is a program executing
on behalf of the work request from the perspective of the work manager. This does not mean that the
program is active from the BCP's perspective.
- ACTIVE APPL
- The active application state sample percentage in contrast to the active subsystem state sample
percentage. This allows a subsystem to differentiate between work requests processed by the
subsystem itself (ACTIVE SUB) and work requests processed by an application invoked by the
subsystem.
- READY
- The ready state sample percentage. Ready indicates that there is a program ready to execute on
behalf of the work request described by the monitoring environment, but the work manager has given
priority to another work request.
- IDLE
- The idle state sample percentage. Idle indicates that no work request (or transaction) is
allowed to run.
|
STATE SAMPLES BREAKDOWN (%) - WAITING FOR |
STATE SAMPLES BREAKDOWN (%) - continued. This category presents up to fifteen named delay
reason states having the highest non-zero values. These are sorted by the sum of the BTE and EXE
rows in each column. The values of the remaining less important states will be accumulated, if
applicable, and presented as delay reason OTHR.
Here is a list explaining the delay reasons.
- LOCK
- waiting for lock.
- I/O
- waiting for I/O indicates that the work manager is waiting on an activity related to an I/O
request. This may either be an actual I/O operation or some function associated with an I/O
request.
- CONV
- waiting for conversation could have been used in conjunction with IWMMSWCH to identify where the
target is located.
- DIST
- waiting for distributed request indicates at a high level that some function or data must be
routed prior to resumption of the work request. This is to be contrasted with 'waiting on
conversation', which is a low level view of the precise resource that is needed. A distributed
request could involve 'waiting on conversation' as part of its processing.
- LOCL
- waiting for a session to be established locally, for example, on the current MVS image.
- SYSP
- waiting for a session to be established somewhere in the sysplex.
- REMT
- waiting for a session to be established somewhere in the network.
- TIME
- waiting for timer.
- LTCH
- waiting for a latch.
- PROD
- waiting for another product.
- MISC
- waiting for unidentified resource, possibly among another specific category, but which may not
be readily determined.
- SSLT
- waiting for an SSL thread.
- REGT
- waiting for a regular thread.
- WORK
- waiting for registration to a work table.
- BPMI
- waiting for I/O resulting from a Db2 buffer pool miss.
- TYPn/TYnn
- The generic delay state defined by the subsystem (possible values: TYP1 - TYP9 and TY10 - TY15).
If the subsystem uses the WLM service IWM4MGDD (Define Descriptions for Generic Delay States) to
provide a description for a generic delay state, RMF displays a legend with the delay state
description. If the subsystem did not use the IWM4MGDD service, the legend is omitted. For further
explanation of the generic delay state types please refer to the subsystem documentation.
|
STATE SWITCHED SAMPL(%) |
Subsystem state samples - continued
- LOCAL
- State representing transactions for which there are logical continuations on this MVS image.
Subsystems might set this state when they function ship a transaction to another component within
the same MVS image.
- SYSPL
- State representing transactions for which there are logical continuations on another MVS image
in the sysplex. Subsystems might set this state when they function ship a transaction to another
component on another image in the sysplex.
- REMOT
- State representing transactions for which there are logical continuations somewhere within the
network. Subsystems might set this state when they function ship a transaction to another component
within the network.
|
Service or Report Class period: goal and actual values
If measurement data for systems in a sysplex is available, this section starts with an *ALL line
showing the average or cumulative values for the sysplex. The *ALL line is followed by one line for
each system.
|
GOAL |
This line shows the goal specified in the WLM service policy for a service class period. For
information about available goal types refer to z/OS MVS Planning: Workload Management. In a report for a homogeneous report class period, the
goal of the corresponding service class period is printed. For heterogeneous report class periods,
N/A is printed.
|
VELOCITY MIGRATION |
The following two values are only provided for periods with an execution velocity goal:
- I/O MGMT
- I/O Priority Management
Value of achieved execution velocity including I/O using and delay
samples.
If WLM I/O delay management is enabled in the service definition, this value matches EX VEL%.
Otherwise, this is the value that would be observed if WLM I/O management were enabled and no other
changes that would affect the execution velocity calculation were made.
You see your current definition in the Service Policy page (see Figure 1).
- INIT MGMT
- Initiator Management
Value of achieved execution velocity including batch initiator delay
samples.
If WLM batch initiator management is enabled in the service definition, this value matches EX
VEL%. Otherwise, this is the value that would be observed if WLM batch initiator management were
enabled and no other changes that would affect the execution velocity calculation were made.
You find a description of these delays in this table for the field TRANS.-TIME.
|
RESPONSE TIME |
This column either shows:
- for an AVG response time goal: the measured average response times
- for a percentile response time goal: the percentages of the transactions that met the response
time goal
- for an execution velocity goal, a system or a discretionary goal:
N/A
- for heterogeneous report class periods:
N/A
|
EX VEL % |
The execution velocity measures the portion of the acceptable processor and storage delays
relative to the total execution time. For details about the execution
velocity, see Common Monitor III report measurements. |
PERF INDX |
The performance index for a period represents how close a period came to reaching the goal
(PI is 1.0 if goal is reached), and how much this period suffered versus its goal. See Table 1 for more details about the performance index. |
AVG ADRSP |
Average number of address spaces and enclaves that contributed delay and using samples to
this class. |
EXEC USING% |
The following using samples are measured as percentages of the total samples:
- CPU
- Standard CP using samples. This value includes using samples of zAAP
and zIIP work executing on general
purpose processors (standard CPs).
- AAP
- zAAP using samples.
- IIP
- zIIP using samples.
- I/O
- I/O using samples.
Note:
- Use the APCUSGP (AAP on CP Using%) overview condition to retrieve the using samples of zAAP work
executing on general purpose processors (standard CPs).
- Use the IPCUSGP (IIP on CP Using%) overview condition to retrieve the using samples of zIIP work
executing on standard CPs.
|
EXEC DELAYS % |
General execution delays included in TOT (total). Each dispatchable unit sampled can increase
one of the CPU or paging delay samples. Besides the TOT value, only the seven highest values
contributing to TOT will be shown. The remaining less important values will be accumulated and
presented as OTH.
- TOT
- Total delay used by SRM in its execution velocity calculation.
- CPU
- CPU delay. A TCB or SRB is waiting to be dispatched (other than the first in-line behind
sampler), or a TCB is waiting for a LOCAL lock.
- AAP
- zAAP-eligible work is delayed because it is waiting for a processor that can run zAAP work.
- IIP
- zIIP-eligible work is delayed because it is waiting for a processor that can run zIIP work.
- I/O
- I/O delay. A TCB or SRB has initiated an I/O request that is delayed obtaining a path to the
device. This includes IOSQ and Q+PEND components (see CONN for a
description).
Note: It depends on the definition in WLM whether this value is part of the TOTAL
value or not, by default it is not contained in TOTAL.
- CAP
- CPU capping delay. A TCB or SRB is marked non-dispatchable because
- a resource group maximum is being enforced
- or because of discretionary goal management. That is, if certain types of work are overachieving
their goals, that work may be capped so that the resources may be diverted to run discretionary work
(see also section 'Using Discretionary Goals' in z/OS MVS Planning: Workload Management).
This value is NOT part of the CPU delay.
- SIN
- Swap-In delay. Swap-In has started but not completed.
- MPL
- MPL delay. Ready but swap-in has not started.
- Q MPL
- Queue MPL - work is waiting for a server address space or batch initiator.
- SRV PRV
- Private area paging delay for a server address space.
- SRV VIO
- VIO paging delay for a server address space.
- SRV SHS
- Hiperspace paging delay for a server address space.
- SRV SIN
- Swap-in delay for a server address space.
- SRV MPL
- MPL delay for a server address space.
- AUX PRV
- Auxiliary paging from private.
- AUX COM
- Auxiliary paging from common.
- AUX XME
- Auxiliary paging from cross memory.
- AUX VIO
- Auxiliary paging from VIO.
- AUX SHS
- Auxiliary paging from standard hiperspaces.
- AUX EHS
- Auxiliary paging from ESO hiperspaces (a page being read was not in the ESO hiperspace, it has
to be read from DASD by the program managing the hiperspace).
|
USING% |
Percentage of using states:
- CRY
- Crypto using state — a TCB or SRB was found to be
using
an adjunct processor (AP).
- CNT
- Contention using state - work is holding resources.
|
DELAY % |
The following states are NOT included in the TOTAL EXECUTION DELAYS.
- UNK
- State is unknown. The address space or enclave was not found to be using or delayed for any
sampled resource, but z/OS has not been notified that it is idle.
- IDL
- Idle state. Work is in STIMER wait, TSO terminal wait, APPC wait, OMVS input or output wait, or
an initiator is waiting for work.
- CRY
- Crypto delay state — a TCB or SRB was found to be waiting for
an AP or a processor feature queue.
- CNT
- Contention delay state - work is waiting for resources.
|
% QUI |
Quiesce state. Some work in this period has been RESET with the QUIESCE keyword. This is the
percentage of address spaces and enclaves quiesced during the reporting interval. |
RESPONSE TIME DISTRIBUTION (for service/report class periods with a response time goal
only) |
WLM maintains counts of how many transactions were completed within a particular time. The
response time goal defined for each service class period is split into 14 response
time buckets where:
- bucket 1 covers the gap from 0 to half the goal
- buckets 2 to 11 cover the gap between half the goal to 1.5 times the goal evenly divided
- bucket 12 covers two times the goal
- bucket 13 covers four times the goal
- bucket 14 covers the gap from four times the goal to infinity
The chart presents the sysplex-wide view on the
- number of total (ended) transactions,
- response time,
- number of total (ended) transactions in percent,
- and a graphical illustration of the percentage.
- TIME
- Response time associated to this bucket.
- # TRANSACTIONS
- Number of transactions that completed for this period.
- CUM TOTAL
- Cumulative number of transactions so far
- IN BUCKET
- Number of transactions in this bucket
- % TRANSACTIONS
- Percentage
- CUM TOTAL
- Cumulative percentage of transactions so far
- IN BUCKET
- Percentage of transactions associated to the bucket
- Percent Scale
- Graphical presentation of each bucket
|
RESPONSE TIME DISTRIBUTIONS (for service/report class periods with an execution velocity goal
only) |
WLM maintains counts of how many transactions were completed within a particular time. The
average of all response times (midpoint) measured for each service/report class period
is split into 14 response time buckets where:
- bucket 1 covers the gap from 0 to half the midpoint
- buckets 2 to 11 cover the gap between half to 1.5 times the midpoint evenly divided
- bucket 12 covers two times the midpoint
- bucket 13 covers four times the midpoint
- bucket 14 covers the gap from four times the midpoint to infinity
For each system in the sysplex, the report presents a tabular representation of:
- the number of total (ended) transactions
- the response time
- the number of total (ended) transactions in percent.
- SYSTEM
- System Name
- INTERVAL
- Measurement interval (Time since last midpoint change)
- MRT CHANGES
- Number of midpoint changes during the SMF interval
- TIME
- Response time associated to this bucket
- # TRANSACTIONS
- Number of transactions that completed for this period
- CUM TOTAL
- Cumulative number of transactions so far
- IN BUCKET
- Number of transactions in this bucket
- % TRANSACTIONS
- Percentage
- CUM TOTAL
- Cumulative percentage of transactions so far
- IN BUCKET
- Percentage of transactions associated to the bucket
|