Service Level Analysis situations

The predefined situations in this category monitor performance, response time, and transaction for potential trouble spots. A description and formula for each of these situations follow.

Note:
All the attribute names in this group are prefixed with CICSplex_Service_Class_Analysis. except where stated.
CICSplex_delay_in_Database
Service Level failure with excessive time spent in database.

Monitors for excessive transaction response time that is attributed to database performance. There are multiple database types (DB2, Adabas, DLI, Datacom, and so forth) that can potentially cause CICSplex delays. CICSplex_delay_in_Database monitors all database types. This situation uses attributes from the CICSplex Service Class_Analysis group. This situation is distributed automatically to the hub server; distribute it to all servers where an OMEGAMON XE for CICSplex is running.

Use the historical transaction display function of OMEGAMON XE for CICS on z/OS to locate the transactions that are responsible for the service level failure. Determine whether the large amount of time spent waiting on database activity is a consequence of either application changes or alterations in the database configuration. If the transaction behavior is found to be normal, consider updating the service class definitions so that they will accurately reflect anticipated response times.

Formula

If the value of Performance_Index is greater than 1 
and (the value of Percent_of_time_waiting_on_Adabas is greater than 90 
 or the value of Percent_of_time_waiting_on_Datacom is greater than 90 
 or the value of Percent_of_time_waiting_on_DB2 is greater than 90 
 or the value of Percent_of_time_waiting_on_DLI is greater than 90 
 or the value of Percent_of_time_waiting_on_IDMS is greater than 90 
 or the value of Percent_of_time_waiting_on_SUPRA is greater than 90
then situation CICSplex_delay_in_Database is true.
CICSplex_delay_in_MQSeries
Service Level failure with excessive time spent in WebSphere MQ.

Monitors the effect of an MQ connection on a service class. This situation evaluates the percentage of time spent waiting for an MQ connection in relation to the performance index for a service class. It alerts the user if the performance index for a service class exceeds 1.0 and the percentage of time spent waiting for an MQ connection exceeds 90%. This situation uses attributes from the CICSplex Service Class_Analysis group. All situations for Service Level Analysis attributes need to be distributed to *HUB. Do not distribute these situations to CICS nodes.

Use the historical transaction display function of OMEGAMON XE for CICS on z/OS to locate the transactions that are responsible for the service level failure. Determine whether the large amount of time spent waiting on message-queuing activity is a consequence of either application changes or alterations to the WebSphere MQ system. If the transaction behavior is found to be normal, consider updating the service class definitions so that they will accurately reflect anticipated response times.

Formula

If the value of Performance_Index is greater than 1 
and the value of Percent_of_time_waiting_on_MQ is greater than 90
then situation CICSplex_delay_in_MQSeries is true.
CICSplex_delay_within_CICS
Monitors service-level failures within internal CICS components attributed to time waiting for CICSplex resources. Initially, this situation acknowledges when performance goals have been exceeded. It then looks for time waiting on CICSplex resources including:

Use the historical transaction display function of OMEGAMON XE for CICS on z/OS to locate the transactions that are responsible for the service level failure. The Bottleneck Analysis report can further assist in locating the wait reasons responsible for slow response time. If CPU bound transactions are unable to get the processor resources they need, consider increasing the priority of either the transaction or the CICS region.

This situation uses attributes from the CICSplex Service Class_Analysis group. This situation is distributed automatically to the hub server; distribute it to all servers where an OMEGAMON XE for CICS on z/OS plex is running.

Formula

If the value of Performance_Index is greater than 1 
and (the value of Percent_of_time_waiting_on_Redispatch is greater than 50
 or the value of Percent_of_time_waiting_on_unidentifiable is greater than 50 
 or the value of Percent_of_time_using_CPU is greater than 50)
then situation CICSplex_delay_within_CICS is true.

See CICSplex_CSM5_delay_in_DLI, CICSplex_CSMI_delay_in_FCP, or CICSplex_delay_in_Database for predefined situations that monitor response time delays in database processing.

CICSplex_CSMI_delay_in_FCP
Service Level failure in the CSMI service class with excessive time in file control.

Use the historical transaction display function of OMEGAMON XE for CICS on z/OS to locate the mirror transactions that are responsible for the service level failure. Determine whether the large amount of time waiting on File Control is a consequence of either application changes or alterations in the CICS file configuration. If the transaction behavior is found to be normal, consider updating the service class definitions so that they will accurately reflect anticipated response times.

Formula

If the value of Service_Class_Name equals CSMI 
and (the value of 200_%_of_Goal_Transaction_Count is greater than 1 
 or the value of Performance_Index is greater than 1 
 or the value of Greater_than_400_%_of_Goal_Transaction_Count is greater than 0 
 or the value of 400_%_of_Goal_Transaction_Count is greater than 1 
and the value of Percent_of_time_waiting_on_File_Control is greater than or equal to 50
then situation CICSplex_CSMI_delay_in_FCP is true.
CICSplex_Function_Ship_delays
Service-level failure in the CSM* service classes in any function-shipping request.

Use the historical transaction display function of OMEGAMON XE for CICS on z/OS to locate the mirror transactions that are responsible for the service level failure. The Bottleneck Analysis report can further assist in locating the wait reasons responsible for slow response time. If CPU limitations are found to be the root cause, consider increasing the priority of the region to prevent File Owning Regions from adversely affecting Application Owning Regions that use function shipping.

Formula

If the value of Service_Class_Name equals "1,CSM"
and the value of Performance_Index is greater than 1
and (the value of Percent_of_time_using_CPU is greater than 50 
 or the value of Percent_of_time_waiting_on_TS_IO is greater than 50 
 or the value of Percent_of_time_waiting_on_Redispatch is greater than 50 
 or the value of Percent_of_time_waiting_on_DLI is greater than 50 
 or the value of Percent_of_time_waiting_on_Journal_Control is greater than 50)
then situation CICSplex_Function_Ship_delays is true.
CICSplex_Performance_Index
Service Level failure: Monitors a CICSplex to see whether it meets current service levels as measured by the performance index. This situation will evaluate true for monitored CICSplexes that fail to meet their specified service levels. This situation uses attributes from the CICSplex Service Class_Analysis group. This situation is distributed automatically to the hub server; distribute it to all servers where a OMEGAMON XE for CICS on z/OS agent is running.

Use the historical transaction display function of OMEGAMON XE for CICS on z/OS to locate the transactions that are responsible for the service level failure. The breakdown of time statistics in the historical transaction detail can identify the component that is causing the response time to be excessive. If the transaction behavior is found to be normal, consider updating the service class definitions so that they will accurately reflect anticipated response times.

Formula

If the value of Performance_Index is greater than 1 
then situation CICSplex_Performance_Index is true.
Note:
This situation is started automatically as part of Tivoli Enterprise Monitoring Server startup. If you do not want this situation to start automatically, open the Settings notebook for this situation to the Properties notebook. Unmark Activate at Start Up, then click OK.
CICSplex_Sympathy_Degradation
Service Level failure: excessive time spent in MRO or ISC: Alerts you before service-class throughput degradation escalates to a more serious problem. However, if degradation continues, then the CICSRegion_Sympathy_Sickness predefined situation alerts you to a CICS failure or reduction in performance because of symptoms in an adjacent region. This situation uses attributes from the CICSplex Service Class_Analysis group. This situation is distributed automatically to the hub server; distribute it to all servers where a OMEGAMON XE for CICS on z/OS is running.

Use the historical transaction display function of OMEGAMON XE for CICS on z/OS to locate the transactions that are responsible for the service level failure. Determine whether the large amount of time spent waiting on MRO is a consequence of a poorly performing remote region or a problem with the connections displayed in the Intercommunication Summary workspace. If the transaction behavior is found to be normal, consider updating the service class definitions so that they will accurately reflect expected response times.

Formula

If (the value of Average_Response_Time is greater than 5 
 or the value of Bucket_G400 is greater than 0 
 or the value of Bucket_200 is greater than 1 
 or the value of Performance_Index is greater than 1 
 or the value of Bucket_400 is greater than 1) 
and the value of Percent_of_time_waiting_on_MRO is greater than or equal to 50 
then situation CICSplex_Sympathy_Degradation is true.
CICSplex_Service_Class_Deleted
z/OS or OMEGAMON XE for CICS on z/OS service-class definition not available.

Formula

If the value of Status equals DELETED
then situation CICSplex_Service_Class_Deleted is true.
CICSplex_CSM5_delay_in_DLI
Monitors the performance of function-shipped remote DL/I (CSM5) requests. It distinguishes between DL/I and non-DL/I problems and alerts the user if more than 50% of the transaction response is attributed to DLI processing.

If service levels for the CSM5 transaction group are not met or response time spikes are detected, CICSplex_CSM5_delay_in_DLI rolls issues an alert.

This situation is distributed automatically to the hub server; distribute it to all servers where an OMEGAMON XE for CICS on z/OS is running. This situation uses attributes from the CICSplex Service Class Analysis group.

Formula

If the value of Service_Class_Name equals "CSM5" 
and (the value of Bucket_200 is greater than 1 
 or the value of Performance_Index is greater than 1 
 or the value of Bucket_G400 is greater than 0 
 or the value of Bucket_400is greater than 1) 
and the value of Percent_of_time_waiting_on_DLI is greater than 50 
then situation CICSplex_CSM5_delay_in_DLI is true.
Note:
CICSplex_CSMI_delay_in_FCP addresses a similar condition with CSMI File Control Processing.
CICSplex_CSMI_Delay_in_FCP
Monitors the performance of function-shipped remote FCP or File Control (CSMI) requests. It distinguishes between FCP and non-FCP problems and alerts the user if more than 50% of the transaction response is attributed to FCP processing.

If service levels for the CSMI transaction group are not met or response time spikes are detected, CICSplex_CSMI_delay_in_FCP issues an alert. This situation uses attributes from the CICSplex Service Class_Analysis group. This situation is distributed automatically to the hub server; distribute it to all servers where an OMEGAMON XE for CICS on z/OSplex is running.

Note:
CICSplex_CSM5_delay_in_DLI addresses a similar condition with CSM5 DL/I database access.

Formula

If the value of Service_Class_Name equals "CSMI" 
and (the value of Bucket_G400 is greater than 0 
 or the value of Bucket_400 is greater than 1 
 or the value of Performance_Index is greater than 1 
 or the value of Bucket_200 is greater than 1) 
and (the value of Percent_of_time_waiting_on_File_Control 
 is greater than or equal to 50) 
then situation CICSplex_CSMI_delay_in_FCP is true.
CICSplex_Function_Ship_Delays
Monitors function shipping performance for CSMI, CSM1, CSM2, CSM3 and CSM5 function shipping types. This situation uses attributes from the CICSplex Service Class_Analysis group.
Note:
See CICSplex_CSM5_delay_in_DLI and CICSplex_CSMI_delay_in_FCP for similar predefined situations. This situation is distributed automatically to the hub server; distribute it to all servers where an OMEGAMON XE for CICSplex is running.

Formula

If the value of Service_Class_Name is greater than "1,CSM" 
and the value of Performance_Index is greater than 1 
and (the value of Percent_of_time_using_CPU is greater than 50 
 or the value of Percent_of_time_waiting_on_TS_IO is greater than 50 
 or the value of Percent_of_time_waiting_on_Redispatch is greater than 50 
 or the value of Percent_of_time_waiting_on_DLI is greater than 50 
 or the value of Percent_of_time_waiting_on_Journal_Control is greater than 50) 
then situation CICSplex_Function_Ship_delays is true.

See also the CICSplex Service Class Analysis attribute group.