Determining the values for the monitoring options in IBM Z OMEGAMON for CICS
Use the following tables to determine the values to specify for the monitoring options. The tables list the groups of tasks and shows the options available for that task.
- Global Options
- Startup Controls
- Exception Analysis Intervals
- Database Collection
- Program Tracking
- Bottleneck Options
- User Excluded Transactions
- User Event Monitoring
- CPU Monitoring
- Group Definitions
- Group Elements
- Interval Collector
- Online Viewer
- Resource Limiting
- Resource Limiting Interval
- Response Time collection
- Dedicated Sessions
- Service Level Analysis
- Enabling outbound API request monitoring
Global Options
Parameter to use | Definition | Default |
---|---|---|
CONVERSE_TIME | Enables you to include or exclude terminal I/O and/or IRC wait time in calculating the total response times for conversational transactions. This value enables MRO users to more completely isolate total transaction response times from I/O-related wait time. Note: CONVERSE_TIME affects only OMEGAMON for CICS data. It does not adjust the
response time collected by CICS® monitoring.
CONVERSE_TIME accepts one or two parameters. The default values for the parameters are indicated.
Note: For RUNTIME and ELAPSED, this global option impacts statistics and functionality. If NOIOWAIT
is enabled (which it is, by default), a task running workload for RUNTIME testing might show lower
values in statistics and warning, but a kill will be performed anyway.
|
NOIOWAIT,IRWAIT |
ETE_INTERVAL | Specifies the ETE response history sampling interval in minutes. Values range from 1-7200 (5 days). | 1 |
ETE_FREQUENCY | Determines how often End-to-End (ETE) response time monitoring occurs for each LU monitored. ETE_FREQUENCY=n specifies that one count takes place every n occurrences of LU activity (1-255). | 1 |
MAX_GROUPS | Specifies the maximum number of groups (1-30) that you can define for internal bottleneck collection, the response time collector, and the interval record collector. | 30 |
MAX_IDS | Specifies the maximum number (1-2000) of transaction IDs, program names, LUs, and terminal IDs for internal bottleneck collection, the response time collector, and the interval record collector. | 50 |
SHUT_OPTIONS | Specifies whether IBM Z OMEGAMON for CICS will purge conversational tasks waiting on terminal I/O when a non-immediate shutdown of CICS is performed. KOCOME00 must be present in the PLT as a shutdown entry. Possible values are:
You can control the purging of tasks during CICS shutdown through the CICS Shutdown Purge Option of the IBM Z OMEGAMON for CICS Classic (3270) interface (fastpath O.Q). |
NOPURGE |
XM_BUFFER_RECORDS | Specifies the number of records (1-2,147,483,648) in the buffer that IBM Z OMEGAMON for CICS uses to accumulate transaction records within CICS before they are sent to the response time collector and the task history collector. This amount should correspond to the average number of transactions within a 2-second period. The default value should be sufficient unless you see the message "OC0014 XMCR communication buffer has wrapped for ..." in the CICS SYSLOG. | 100 |
Startup Controls
Parameter to use | Definition | Default |
---|---|---|
BOTTLENECK_ANALYSIS | Specifies whether or not internal bottleneck collection starts automatically. Possible values are:
Note: NOAUTO cannot be specified if you also specify
INTERVAL_RECORDING=AUTO.
|
AUTO |
CPU_THRESHOLDING | Determines whether or not the MAXR exception is tripped when CPU consumption exceeds the MAXCPU threshold. The MAXR exception takes effect only if you use MAXCPU to specify a CPU consumption threshold greater than 0. In addition, CICS version 3.x users must specify CPU=YES in the MCT to instruct CICS to monitor CPU consumption. Possible values are:
Note: Once a transaction has tripped the MAXR exception, that transaction will continue to appear on
the Tasks Exceeding Global CPU Time Limit panel of the OMEGAMON for CICS (3270) interface, even if you increase
the CPU consumption threshold. Use the more powerful features of resource limiting instead
of MAXR.
|
AUTO |
DB2_CLOCKS_AND_COUNTERS | Specifies whether the DB2® data collector starts up automatically. DB2 collector accumulates elapsed times and counts of DB2 commands for each CICS transaction. Possible
values are:
If you specify NOAUTO, you can dynamically start and stop DB2 collection through the Control Database Collectors panel of the OMEGAMON for CICS (3270) interface (fastpath O.P). |
AUTO |
DLI_CLOCKS_AND_COUNTERS | Specifies whether the DL/I data collector starts automatically. The DL/I collector accumulates elapsed times and counts of DL/I commands for each CICS transaction. Possible
values are:
Specifying AUTO or NOAUTO enables the XDLIPRE and XDLIPOST global user exits. You can dynamically start and stop DL/I collection through the Control Database Collectors panel of the OMEGAMON for CICS (3270) interface (fastpath O.P). If you specify NO, however, you cannot activate DL/I collection dynamically. |
AUTO |
INTERVAL_RECORDING | Specifies whether the interval record collector starts automatically. Possible values are:
You can start the interval record collector through the Control Interval Recording Collector panel of the OMEGAMON for CICS (3270) interface (fastpath O.K). Note: AUTO
requires that BOTTLENECK_ANALYSIS=AUTO also be specified in the STARTUP_CONTROL
component.
|
NOAUTO |
ONLINE_DATA_VIEWING | Specifies whether the task history collector starts up automatically. Possible values are:
You can start the task history collector through the Start Online Historical Transaction Viewing Collector panel in the OMEGAMON for CICS (3270) interface (fastpath O.F). |
AUTO |
RESOURCE_LIMITING | Specifies whether OMEGAMON for CICS automatically monitors tasks for resource limiting once it is initialized. Possible
values are:
You can start resource limiting through the START RESOURCE LIMITING panel of the OMEGAMON for CICS (3270) interface (fastpath O.R). |
NOAUTO |
RESOURCE_LIMITING_MSG_DEST | Specifies the destination for RLIM messages OC8902 and OC8903. Following is a list of possible resource message destination values:
The current settings may be viewed on the OMEGAMON for CICS (3270) system's RESOURCE LIMITING STATUS panel (fastpath O.T). |
TDQ |
RESOURCE_LIMITING_SYSTEM_TASKS |
This parameter allows you to use the WARN thresholds specified in Resource Limiting to alert you to any system tasks which exceed the WARN threshold. It also allows you to KILL tasks using IBM-provided programs, these might include CEMT or CICS mirror transactions. CICS transaction manager determines which transactions are system tasks. Possible values are:
|
NO |
RESOURCE_LIMITING_ABEND_CANCEL | Specifies whether the CANCEL option is used by RLIM when KILLing a transaction Possible values are:
Note: RLIM will attempt to KILL a task only once.
If a task continues to process and exceeds other limits RLIM will
take no action for this task.
|
YES |
RESOURCE_LIMITING_TRACE_WARNED | Specifies whether transactions that are warned by RLIM will have Application Trace activated for the task automatically Possible
values are:
|
YES |
ENABLE_APPLICATION_TRACE | Specifies whether OMEGAMON for CICS will be able to record Application Trace for transactions running in CICS. Possible values are:
You can enable Application Trace through the Application Trace Facility Status panel of the OMEGAMON for CICS (3270) interface (fastpath O.W). |
NOAUTO |
ENABLE_FILE_CONTROL_EXIT | Specifies whether global user exits XFCFRIN and XFCFROUT will be initialized. These exits enable OMEGAMON for CICS to collect file level data for transactions which access VSAM files. Possible
values are:
|
NO |
RESPONSE_TIME_ANALYSIS | Specifies whether the response time collector starts up automatically. Possible values are:
You can start the response time collector during a session using the START RESPONSE TIME MONITOR panel of the OMEGAMON for CICS (3270) interface (fastpath O.A). |
AUTO |
AUTO_RESTART | Specifies whether, after the OMEGAMON for CICS address space is restarted, OMEGAMON for CICS is not restarted in the CICS region Possible values
are:
|
NO |
HISTORY_CATCHUP_TIME | Specifies the maximum time in seconds (0-300) after a restart of the OMEGAMON for CICS address space that OMEGAMON will wait for the Online Data Viewing component to retrieve the data in the OMEGAMON buffers before starting the recycle of the code in the CICS region. If a value of 0 is coded (the default), no attempt is made to start history prior to recycling the OMEGAMON code in the CICS region. | 0 |
TRANSACTION_TRACKING | Specifies whether the Transaction Tracking API will be initialized when the OMEGAMON for CICSaddress space is started. Possible
values are:
|
NOAUTO |
Exception Analysis Intervals
Parameter to use | Definition | Default |
---|---|---|
VSAM_MONITORING_INTERVAL | Specifies the interval of time for which statistics concerning VSAM CA splits, CI splits and extents are collected. Possible values for this field are 1-1440 minutes. A value of zero means statistics are collected and evaluated every time the Region Status or VSAM panels are refreshed. | 15 minutes |
DUMPS_MONITORING_ INTERVAL | Specifies the interval of time for which statistics concerning system dumps and transaction dumps are collected. Possible values for this field are integers from 1-60 minutes. | 15 minutes |
VIOLATIONS_MONITORING_ INTERVAL | Specifies the interval of time for which statistics concerning storage violations are collected. Possible values for this field are 1-60 minutes. | 15 minutes |
Database Collection
In this section of the Global Data Areas values, you can select database products to be monitored. You can specify ADABAS, Db2, DATACOM, DLI, IDMS, SUPRA, VSAM, MQ, or USREVNT1 for monitoring to get file I/O statistics by transaction.
VSAM statistics are derived from only EXEC-level file control requests.
To capture DL/I statistics at OMEGAMON for CICS startup, you must specify DLI_CLOCKS_AND_COUNTERS=AUTO on the STARTUP_CONTROL component. DL/I statistics are from EXEC-level or CALL-level requests. For local or remote PSBes, the statistics are captured by DBD name. For DBCTL, the statistics are captured by PSB name.
For ADABAS, CA-DATACOM, IDMS, and SUPRA, you must carry out the installation process as described in Installing third-party support
If you wish to monitor an event other than the supported Files or Databases, specify the USREVNT1 sub-component.
- DLI=YES will enable XDLIPRE and XDLIPOST.
- MQ=YES will enable XRMIOUT.
- TRACE=YES will enable XRMIOUT and XPCABND.
Parameter to use | Definition | Default |
---|---|---|
AUTO_START | Specifies that collection starts up automatically. | YES |
ONDV_WRITE | Specifies that statistics are to be available for the task history collector. | YES |
SMF_WRITE |
Specifies whether to write statistics to SMF for batch reporting. Possible values are:
TRACE data cannot be written to SMF. If you select this option the SMF_WRITE=YES parameter card will be discarded. |
NO |
DETAIL |
Specifies report formatting option for ADABAS clock and counter statistics. Possible values are:
|
NO |
Parameter to use | Definition | Default |
---|---|---|
EVENT | Defines a 1-8 character name. This operand is used only when USREVNT1 is specified. It denotes an in-house database name, a user program name, or any other user event. | None |
FUNCTION | Defines a list of 1 to 10 names. Each name represents a pair of buckets used to hold the clock and count statistics related to the previously defined EVENT. A FUNCTION name must be between 1 and 8 characters long. | None |
EVENT=MYDBASE
FUNCTION=(BROWSE,ADD,UPDATE,DELETE)
Program Tracking
Parameter to use | Definition | Default |
---|---|---|
ENABLE | Specifies whether Program Tracking is enabled. AUTO: Start Program Tracking automatically. NOAUTO: Do not automatically start Program Tracking. | NOAUTO |
AGGREGATE | This option will allow you to see the metrics for program usage across the entire CICS region. These metrics are accumulated as of when the task ends. Looking at programs used within a CICS region, the Used tab will show the collective metrics for all programs which have had data collected since the region started or program manager statistics were reset. | NOAUTO |
WRITE_TO_HISTORY | This option will record the collected program data for a task to Task History. This will provide the Programs tab when looking at the details for a task. This tab will provide information about all programs used during the time covered by the history record. | YES |
SMF | This option will record the collected information to SMF. The data can then be processed offline by utilities such as IBM CICS Performance Analyzer. | NO |
Bottleneck Options
Parameter to use | Definition | Default |
---|---|---|
CLEAR_INTERVAL_LONG | Specifies (in minutes) the default CLEAR interval for long-term accumulators (0-999). To disable this automatic reset, specify 0. | 30 |
CLEAR_INTERVAL_SHORT | Specifies (in minutes) the default CLEAR interval for short-term accumulators (0-999). Zero (0) resets the short-term accumulators every internal bottleneck collection cycle. | 10 |
SAMPLE_INTERVAL | Specifies (in tenths of a second) the sample interval for internal bottleneck collection (1-99). Note: The
relationship between SAMPLE_INTERVAL and CPU usage is not linear.
For example, changing the interval from 2 to 4 seconds does not yield
a 50% reduction.
|
23 |
VARIABLE_BUCKETS | Specifies the maximum number of variable buckets that internal bottleneck collection can allocate. The value can range from 0-32767. Whenever the allocated storage overflows, the Bottlenecks panel in the OMEGAMON for CICS (3270) interface displays a new line that shows the wait reason sub-ID of *OVERFLW* and also the percentage of waits that used the overflow area. Note: Use
the following formula to compute the amount of storage required for
each variable bucketset.
Where X is
the maximum number of groups coded on the MAX_GROUPS operand of the
GLOBAL_OPTIONS component. |
1000 |
EXCLUDE_SYSTEM_TASKS | Specifies whether system tasks waiting on resources are excluded from Bottleneck Analysis. Possible values
are:
|
NO |
EXCLUDED_TRANS | Specifies transactions to be excluded from Bottleneck Analysis. This can be the actual 1- to 4-character transaction ID(s) or a generic specification using wildcard characters. You can use the question mark (?) wildcard character for single-character substitution and the asterisk character (*) as a trailing character only. The
following are examples of wildcarding:
The following transaction IDs are always excluded from
Bottleneck Analysis:
Note: If the list of transactions will not fit on one line,
use additional EXCLUDED_TRANS statements.
|
None |
BOTTLENECK_ANALYSIS | Specifies whether internal bottleneck collection starts automatically. Possible values are:
Note: NOAUTO cannot be specified if you also specify
INTERVAL_RECORDING=AUTO.
|
AUTO |
User Excluded Transactions
To simplify your user-defined situations, you can specify a list of transactions that you want to exclude from being counted toward a situation. This user excluded list of transactions will reside in the OMEGAMON for CICS global data member.
To exclude transactions, specify the USER_EXCLUDED_TRANS
parameter, with the
TranID(s) you want to exclude.
A user-defined situation can then check for the TranID value(s) you specified. If a transaction
has a name that matches the value of USER_EXCLUDED_TRANS
, that transaction will not
be included in the situation. You can use the question mark (?) and asterisk (*) wildcard characters
in the transaction list; the asterisk can only be used as a trailing character.
CPU_EXCLUDED_TRANS
section in the global
data member was repurposed as USER_EXCLUDED_TRANS
for use with situations. If you
refresh your global data member or perform a Reformat (KC2GLBVR Job) , the
CPU_EXCLUDED_TRANS
section will be renamed USER_EXCLUDED_TRANS
.
When the TRAN attribute group formats a row, it will check if the TranID, as displayed, matches an
entry in the USER_EXCLUDED_TRANS
list. If so, the new excluded attribute is set to
YES. If the TranId is not in the list, the excluded attribute is set to NO. There are two other
conditions where the agent cannot make a determination. In these cases, a specific value is used for
the attribute:- If the KOCCI is not active. The attribute value
O
(No_KOCCI) is used. - If the global data member is not loaded. The attribute value
G
(No Global) is used.
Parameter to use | Definition | Default |
---|---|---|
EXCLUDED_TRANS | The transactions you want to exclude from situations.
The exclude list may contain the actual 1- to 4-character transaction ID(s) or a generic specification using wildcard characters. You can use the question mark (?) wildcard character for single-character substitution and the asterisk (*) as a trailing character only. The following are examples of wildcarding:
Excludes the ABCD transaction(s) and all transaction IDs that begin with E.
Excludes all 2-character transaction IDs where the first character is F.
|
None |
User Event Monitoring
Defines and tailors user-event monitoring to OMEGAMON for CICS. OMEGAMON for CICS turns on several global user exits and, if not already activated, CICS CMF performance monitoring. This allows OMEGAMON for CICS to collect additional data about transaction execution.
This data is used by the response time collector, interval record collector, resource limiting, and task history collector; and appears in SMF 112 and SMF 110 records written to SMF from the CICS address space by both OMEGAMON for CICS and CICS. This additional data causes a slight increase in the pathlength of CICS transactions and in the CPU usage of the monitored CICS address space, depending on the transaction load.
OMEGAMON for CICS lets you limit the additional CPU utilization introduced by OMEGAMON for CICS into CICS transaction execution path length.
CICS_CMF_MONITORING can be used to suppress the automatic activation of CICS CMF performance monitoring.
Parameter to use | Definition | Default |
---|---|---|
BASIC_SECTION=OMEGBSC | Specifies the entry name of the event-monitoring point in the MCT for the BASIC section. This name must match the ID operand of the DFHMCT macro. | OMEGBSC |
DB2_SECTION=OMEGDB2 | Specifies the entry name of the event-monitoring point in the MCT for the DB2 section. This name must match the ID operand of the DFHMCT macro. | OMEGDB2 |
DLI_SECTION=OMEGDLI | Specifies the entry name of the event-monitoring point in the MCT for the DLI section. This name must match the ID operand of the DFHMCT macro. | OMEGDLI |
MQ_SECTION | Specifies the entry name of the event-monitoring point in the MCT for the MQ section. This name must match the ID operand of the DFHMCT macro. | CANMQ |
USREVNT1_SECTION | Specifies the entry name of the event-monitoring point in the MCT for the USREVNT1 section. This name must match the ID operand of the DFHMCT macro. | CANUE1 |
CICS_CMF_MONITORING | Suppresses startup CMF performance monitoring only if not already active. Possible values are:
Notes®:
|
YES |
CICS_CMF_WRITE | Determines whether CMF is allowed to generate and write SMF type 110 records. Possible values are:
Note: If you specify CICS_CMF_WRITE=NO and CICS_CMF_MONITORING=YES, OMEGAMON for CICS turns on CMF in order to
collect transaction data about response time and task history, but does not log this data
to SMF (for batch reporting by the historical reporter).
|
YES |
COMPRESS_SMF112_RECORDS | Controls the compression of the OMEGAMON for CICS SMF 112 (database) records during XMNOUT processing. Possible values are:
You can optionally write database performance data to SMF. The OMEGAMON for CICS SMF buffers may contain data for multiple CICS transactions, and will be written to SMF when the buffer fills. Optionally the data can be compressed before the write to SMF occurs. This compression is performed using a standard compression utility, CSRCESRV. |
NO |
MCT_VALIDATION_MESSAGES | OMEGAMON EMPs are required if data is to be written to SMF. This field specifies whether the MCT scan routine will display messages. Possible values are:
|
NO |
CPU Monitoring
Parameter to use | Definition | Default |
---|---|---|
CPU_THRESHOLD | Specifies the CPU-consumption threshold in seconds (0-32,768). The default, 0, means that the CPU-consumption exception (MAXR) will not be tripped. Note: The CPU_THRESHOLDING
operand of the STARTUP_CONTROL component determines whether the MAXR
exception is tripped when the threshold is exceeded. Once a transaction
has tripped the MAXR exception, that transaction will continue to
appear on the Tasks Exceeding Global CPU Time Limit panel, even if
you increase the CPU-consumption threshold.
|
0 |
Group Definitions
Group Definitions define transaction, terminal, and program groups for internal bottleneck collection.
You can also dynamically define groups through the Groups panel of the OMEGAMON for CICS (3270) interface (fastpath,G.), but the changes remain in effect only temporarily.
- You can define as many as 30 groups.
- Each group can contain one or more element names, such as a CICS transaction ID or terminal ID.
- All the elements in a group must be of the same resource type.
- You can define a maximum aggregate of 2000 elements across all groups.
- The same element can appear in multiple groups.
Code one subentry for each group you define. Each entry must have a unique group number and group name.
Parameter to use | Definition | Default |
---|---|---|
GROUP_NUMBER | Specifies the group number (01-30) or the maximum number of groups allowed as specified by the MAX_GROUPS operand of the GLOBAL_OPTIONS component. | 01 |
GROUP_NAME | Specifies the name of this group (1-12 characters). | (TRAN GRP A* ) |
GROUP_TYPE | Specifies the group type (PROG, TERM, or TRAN). | TRAN |
RESPONSE_TIME_THRESHOLD | Specifies average response time thresholds for this group (1-999 tenths of a second) for the interval record collector and the response time displays. Note: OMEGAMON for CICS, uses the value you
specify as the Warning threshold for the response time light. Also, OMEGAMON for CICS doubles the
RESPONSE_TIME_THRESHOLD value to compute the critical threshold for the response time
light.
|
20 |
Group Elements
Group Elements define elements in one or more monitoring groups.
You can also dynamically define elements within groups through the Groups panel of the OMEGAMON for CICS (3270) interface (fastpath G.).
Specifies information for an ID entry. You must specify at least one ID.TERM, TRAN, PROG, LU, and TASKREQ are mutually exclusive. You must code a separate ID for each element that you wish to include in a group.
You can use the PROG, TERM, and TRAN keywords with an asterisk (*) as a wildcard character to represent all possible characters. For example, TERM=L* specifies all terminals beginning with the letter L.
Parameter to use | Definition | Default |
---|---|---|
GRAPH_SCALE | Specifies the height (1-99 seconds) of the response time problems (RSP) graph. This entry is for the response time collector only. | 2 |
TIME_WINDOW | Specifies the selection time window (1-10 minutes) for the RSP dynamic slot analysis (moving time slots). This entry is for the response time collector only. | 10 |
ID |
Specifies the start of the ID entry. You must code a separate ID entry for each element that you wish to include in a GROUP. Possible values are:
Note: TRAN, PROG, TERM, and TASKREQ are mutually exclusive.
|
None |
ELIGIBLE_GROUPS | Specifies the GROUP or GROUPS with which the ID is associated. You can specify from 1 to the maximum number of groups allowed according to the MAX_GROUPS operand of the STARTUP_CONTROL component. | (01) |
TOTAL_RESPONSE_THRESHOLD | Specifies the response time threshold (0-999 tenths of a second) at which this ID appears in the dynamic portion of the RSP display. This entry is for the response time collector only. | 20 |
HOST_RESPONSE_THRESHOLD | Specifies the host response time threshold (0-999 tenths of a second). If an LU exceeds this threshold, the ETE response time displays show the host response time in red and record the condition in the response history graph with a red bar. This operand can be specified only for LU IDs. | None |
NETWORK_RESPONSE_THRESHOLD | Specifies the network response time threshold (0-999 tenths of a second). If an LU exceeds this threshold, the ETE response time displays show the network response time in red and record the condition in the response history graph with a red bar. You can use this operand only for LU IDs. | None |
FIXED_DISPLAY |
Indicates whether the ID should remain on the graph presented by the RSP command in the OMEGAMON for CICS (3270) system, regardless of the response time. Possible values are:
|
NO |
Interval Collector
This set of values defines interval record collector options. After the options are defined in the global data area module, you can control the interval record collector options through the Control Interval Recording Collector panel in the OMEGAMON for CICS (3270) system (fastpath O.K).
Parameter to use | Definition | Default |
---|---|---|
NUMBER_DASD_DEVICES | Specifies the number of internal queue elements for the I/O sub-analysis routine (the number of DASD devices allocated to the CICS region). Valid values are 0-99. A value of 0 indicates that no I/O sub-analysis will be conducted by Interval Recording. | 0 |
RECORDING_INTERVAL | Specifies the interval recording collection interval. This interval may be set to a value from 1-255 minutes. | 1 minute |
Online Viewer
This section defines options for the task history collector. The following values are defined:
Parameter to use | Definition | Default |
---|---|---|
DATA_STORE_TYPE |
Defines the way that data is saved. Possible values are:
Whether you use a data space or linear dataset, the task history collector writes records using a wraparound format. Once the data space or dataset is full, the task history collector resumes writing records from the beginning, overlaying the previous data. |
DSPACE |
DATA_STORE_SIZE | If you specify DSPACE, you must specify the size of the data space in kilobytes. Each transaction record requires about 808 bytes. The maximum size is 2,097,148 kilobytes. | 956 |
RESERVED_SIZE |
Specifies the portion of the data space (from 0 to 50 percent) reserved for file, database clock and counter statistics, and Application Trace records. See the Application Trace feature in the OMEGAMON for CICS. You must specify a number greater than 0 or statistics by filename and/or Application Trace records will not be available from the task history collector. The RESERVED_SIZE operand is not needed when the DATA_STORE_TYPE=FILEOCMP is specified. It will be ignored if it is supplied. File and database clock and counter statistics are always available in the FILEOCMP data store when they are being collected in the CICS region (through use of the DATABASE_COLLECTION component). |
25 |
DATA_STORE_FILE_NAME |
Specifies the name of a VSAM linear dataset when the DATA_STORE_TYPE=FILEOCMP or DATA_STORE_TYPE=AUTO is specified. You can share global data area modules (but not VSAM data sets) across different CICS regions even though a dataset name is specified in the global data area module. To do this, you must insert one wildcard character (*) somewhere in the VSAM dataset name on this operand. When the task history collector is running, the wildcard character is replaced by the job name of the CICS region being monitored. The wildcard character may appear anywhere in the dataset name and should occur only once. |
None |
EXCLUDED_TRANS |
Indicates transactions to exclude from analysis. The list of transactions is limited to 63 transaction IDs. Transactions that start before phase 3 of the PLTPI or end after phase 1 of the PLTSD are not excluded. Any CICS-generated transactions that initiate without going through TRUE processing are not excluded either. The exclude list may contain the actual 1- to 4-character transaction ID(s) or a generic
specification using wildcard characters. You can use the question mark (?) wildcard character for
single-character substitution and the asterisk character (*) as a trailing character only. The
following are examples of wildcarding:
Excludes the
ABCD transaction(s) and all transaction IDs that begin with E. Excludes all 2-character transaction IDs where the first character is
F. Excludes all 4-character transaction IDs starting with G and ending
with H.Note: If the list of transactions will not fit on one line, use additional EXCLUDED_TRANS
statements.
|
None |
Resource Limiting
Resource Limiting defines resources with their respective thresholds and identifies the action to be taken when a resource threshold or limit has been exceeded.
Parameter to use | Definition | Default |
---|---|---|
Resource |
Specifies a resource for limiting usage:
|
None |
EXCLUDED_TRANS |
Specifies one or more transactions to exclude from resource limiting. This can be the actual 1- to 4-character transaction ID(s) or a generic specification using wildcard characters. You can use the question mark (?) wildcard character for single-character substitution and the asterisk character (*) as a trailing character only. The following are examples of wildcarding:
Excludes
the ABCD transaction(s) and all transaction IDs that begin with E Excludes all 2-character transaction IDs where the first character is
F. Excludes all 4-character transaction IDs starting with G and ending
with H.Non-displayable transaction IDs are specified as follows:
Note: Neither KILL_LIMIT nor WARN_LIMIT can be used with EXCLUDED_TRANS.
Exclude CICS system transactions by specifying EXCLUDED_TRANS=C*. If the list of transactions will not fit on one line, use additional EXCLUDED_TRANS statements. |
|
INCLUDED_TRANS | Specifies one or more transactions for resource limiting. This can be the actual 1- to 4-character transaction ID(s) or a generic specification using wildcard characters. You can use the question mark (?) wildcard character for single-character substitution and the asterisk character (*) as a trailing character only. The following are examples of wildcarding: Selects the ABCD transaction(s) and
all transaction IDs that begin with E Selects all 2-character transaction IDs where the first character is
F. Selects all 4-character transaction IDs starting with G and ending
with H.Non-displayable transaction IDs are specified as
follows:
Note: If INCLUDED_TRANS is specified, either KILL_LIMIT or
WARN_LIMIT (or both) must also be specified. See below for information on these
parameters.
If the list of transactions will not fit on one line, use additional INCLUDED_TRANS statements. |
None |
KILL_LIMIT |
Specifies a threshold for limiting resource usage. Transactions that exceed this value are abended. Thresholds can be set at the following values:
|
None |
WARN_LIMIT |
Specifies a threshold for limiting resource usage. OMEGAMON for CICS issues a warning message for transactions that exceed this value. The value must be lower than the KILL_LIMIT if specified. |
None |
A transaction executing under CEDF is not abended. Instead, a message is issued to the terminal and the transaction continues normal execution. OMEGAMON for CICS writes messages, issued as a result of KILL or WARN action, to the CICS Transient Data Queue (ID of CSSL, dataset MSGUSR) or via a WTO.
Transaction selection tips:
Since you might have some overlap in transaction codes if you use both EXCLUDED_TRANS and INCLUDED_TRANS with wildcarding, OMEGAMON for CICS selects a transaction for resource limiting as follows:
- If you do not specify the INCLUDED_TRANS operand or if you specify the INCLUDED_TRANS operand
that matches a specification in the list, OMEGAMON for CICS selects the transaction. If a
transaction ID qualifies for resource limiting under more than one rule, the most specific
INCLUDED_TRANS specification is used. For
example:
Rule 1: Rule 2: ------------------- ------------------- <<DB2>> <<DB2>> INCLUDED_TRANS=TST* INCLUDED_TRANS=TST1 KILL_LIMIT=100 KILL_LIMIT=50 WARN_LIMIT=80 WARN_LIMIT=40
Rule 1: Rule 2:
------------------- -------------------
<<DB2>> <<DB2>>
INCLUDED_TRANS=TST1 EXCLUDED_TRANS=T*
KILL_LIMIT=100
WARN_LIMIT=80
Rule 1: Rule 2:
------------------- -------------------
<<DB2>> <<DB2>>
INCLUDED_TRANS=TST* EXCLUDED_TRANS=TST1
KILL_LIMIT=100
WARN_LIMIT=80
Then transaction TST1 will be excluded from the resource limit.RLIM versus ICVR tips:
The ICVR parameter, available through CICS, and the RLIM CPU option, available through OMEGAMON for CICS, serve two separate functions:
- The ICVR value allows you to set an upper limit for the amount of CPU time that a task may consume within a single dispatch.
- The RLIM CPU limit allows you to set an upper limit on the total amount of CPU time that a task may consume across multiple dispatches.
You may consider RLIM and ICVR to be complementary tools you can use to prevent a task from monopolizing the CPU.
When determining whether or not a task should be abended due to excessive CPU consumption, RLIM compares the value in USRCPUT, DFHTASK_008 in the Transaction Monitoring Area, with the CPU limit specified in the appropriate global data area module, KC2GLBnn.
The CICS Monitoring Facility updates the value in USRCPUT when a task issues a CICS command that suspends the task and causes the dispatcher to be invoked. RLIM CPU usage checking is performed whenever an EXEC CICS request is issued. Additionally, if you have installed support for a third-party database product and are collecting clock and counter statistics for that database type, a request issued to the database will cause the task's CPU usage to be checked.
- ASKTIME
- HANDLE
- ENQ
- DEQ
- FREEMAIN
- RELEASE
- TRACE ON/OFF
- XCTL
In addition, a READ request that can be satisfied from the buffer, as opposed to requiring physical I/O, does not invoke the dispatcher. Other CICS commands may invoke the dispatcher, but only if the request cannot be satisfied immediately and the task is suspended. Therefore, an application that, for instance, repeatedly issues a GETMAIN and FREEMAIN for a small area may not give up control. In such a situation, the task's CPU time value will not be updated and RLIM will not abend the task (see "External task resource limiting," below). Conversely, because the USRCPUT value is updated only at the end of each dispatch, the CICS runaway-task mechanism, controlled by the ICVR parameter in the SIT, does not use it. To determine whether or not to abend the task, CICS refers to a timer that is reset each time the task is redispatched. While dispatched, CICS decrements the timer at regular intervals. If the timer reaches zero, the task is abended with code AICA.
External task resource limiting
When RESOURCE_LIMITING_MSG_DEST is specified as "LOG" or "TDL," RLIM runs an additional program each second which monitors active tasks and, if needed, takes action against a task. This is done only if limits were specified for resource types "CPU" or "ELAPSED" and is intended to handle tasks that are in a long wait or actively using CPU but not issuing EXEC CICS requests. Message KCP8906I is issued and serves as a notification that the task has exceeded the threshold even though RLIM cannot ABEND it. You may use this message to automate actions to notify operators that external action may be required.
Resource Limiting Interval
The RESOURCE_LIMITING_INTERVAL specifies at what interval resource limiting should process user defined rules. Once RLIM is active, the interval settings apply to all resources and their respective rules, hence a global implication. There are two ways to specify an interval. Either, specify a TIME_INTERVAL or specify EXEC_CALLS and/or DB_CALLS.
The values in the following tables can be used.
Parameter to use | Definition | Default |
---|---|---|
TIME_INTERVAL | Defines the amount of time, in units of seconds, which must pass before existing resource limits or rules could be processed. Default value of zero seconds implies that RLIM processing will not be based on time interval. Maximum allowable value is one week or 604800 seconds. | 0 |
EXEC_CALLS | Defines the number of EXEC CICS commands which must be issued by the application program before existing resource limits or rules could be processed. Default value of zero implies that RLIM processing will not be based on the number of EXEC calls. Maximum allowable value is 99999999. | 0 |
DB_CALLS | Defines the number of DataBase commands or EXEC CICS SQL commands which must be issued by the application program before existing resource limits or rules could be processed. Default value of zero implies that RLIM processing will not be based on the number of DataBase or EXEC CICS SQL calls. Maximum allowable value is 99999999. | 0 |
Only those transactions which are started after the intervals were last refreshed, are eligible for processing.
Response Time collection
The RESPONSE_TIME_COLLECTION value specifies the time intervals for the response time collector.
The values in the following table can be used.
Parameter to use | Definition | Default |
---|---|---|
TIME_INTERVALS | Specifies the time intervals for the Average Response Times display. You must enter three values:
These are the three intervals for reporting RTA data. In order to roll the values from one interval to the other, OMEGAMON for CICS requires that "long" be an integer multiple of "mid" and that "mid" be an integer multiple of "short". The values are in minutes within the range 1-999. |
(5,10,30) |
EXCLUDED_TRANS | Specifies transactions to exclude from analysis. The list of transactions is limited to 63 transaction IDs. Transactions that start before phase 3 of the PLTPI or end after phase 1 of the PLTSD are not excluded. Any CICS-generated transactions that initiate without going through TRUE processing are not excluded either. The exclude list may contain the actual 1- to 4-character transaction ID(s) or a generic specification using wildcard characters. You can use the question mark (?) wildcard character for single-character substitution and the asterisk character (*) as a trailing character only. The following are examples of wildcarding: Excludes the ABCD
transaction(s) and all transaction IDs that begin with E. Excludes all 2-character transaction IDs where the first
character is F. Excludes all 4-character transaction IDs starting with G and
ending with H.Note: If a transaction is matched in the EXCLUDED_TRANS list, OMEGAMON for CICS excludes it from analysis,
whether or not it matches a GROUP_ELEMENTS component on either terminal ID, program,
transaction ID, LU name, or TASKREQ.
If the list of transactions will not fit on one line, use additional EXCLUDED_TRANS statements. |
None |
TIME_SLOT | Specifies the time slot definitions (1-48 slots) for the Today's Response Times display. It specifies the start time (first entry) and end time (second entry) of the slot. Enter the values in 24-hour clock format using this syntax:
Where At least one slot must be defined; if more than one slot is defined, the slots must be in ascending order. The ending time value on one slot may be the same as the beginning time value on the next slot, but overlaps are not allowed. (A shared minute is not an overlap.) The default RTA TIME_SLOTs daily configuration is one slot from 0000 to 0800, then contiguous hourly slots until 1700, and one more slot from 1700 to 2400. |
|
Dedicated Sessions
The DEDICATED_SESSIONS values define terminals for OMEGAMON for CICS dedicated mode sessions.
The values in the following tables can be used.
Parameter to use | Definition | Default |
---|---|---|
UNIT | Specifies the beginning of information for an external monitoring session. It is followed by the other operands defining a session. | None |
UNIT_ADDRESS | Specifies the device address of the terminal. | None |
PHYSICAL_ROWS | Specifies the number of physical rows (1-999) on the terminal. | 24 |
LOGICAL_ROWS | Specifies the number of logical rows (1-999). | 99 |
COLUMNS | Specifies the number of columns (80, 132, or 160) on the terminal. | 80 |
USER_PROFILE | Specifies the user profile suffix (any two characters). Any session of OMEGAMON for CICS can use a different profile by specifying a different two-character suffix. | None |
Service Level Analysis
The SERVICE_LEVEL_ANALYSIS value controls whether data will be written to the Service Level Analysis subtask running in the CICS agent address space.
The value in the following tables can be used.
Parameter to use | Definition | Default |
---|---|---|
INCLUDE_REGION | Specifies whether performance data collected by OMEGAMON is sent to the SLA summarization routines running in the OMEGAMON for CICS agent address space and displayed in the Service Level Analysis reports. Possible values are:
|
YES |
Enabling outbound API request monitoring
The OUTBOUND_API_REQUEST_MONITORING parameter, AGGREGATE, controls whether file level monitoring will be enabled, to collect data for outbound API request monitoring. The following table gives the values that can be used.
Parameter to use | Definition | Default |
---|---|---|
AGGREGATE | Specifies whether file level collection for outbound API request
monitoring is enabled.
|
NOAUTO |
TRUE Monitoring
The TRUE_MONITORING values control whether data for Task Related User Exits (TRUE) will be collected, written to task history, and written to SMF so that the data can be processed offline by utilities such as IBM CICS Performance Analyzer.
Parameter to use | Definition | Default |
---|---|---|
ENABLE | Specifies whether TRUE Monitoring is enabled. AUTO: Start TRUE Monitoring automatically. NOAUTO: Do not start TRUE Monitoring automatically. | NOAUTO |
WRITE_TO_HISTORY | This option will record the collected data for a task to Task History. This tab will provide information about all TRUEs used during the time covered by the history record. | YES |
SMF | This option will record the collected information to SMF. The data can then be processed offline by utilities such as IBM CICS Performance Analyzer. | NO |