Besides the above-mentioned global system workload values, additional data is required for detecting delays. To begin with, the corresponding task selection can be carried out using the periodic task statistics (monitoring program PERIODIC-TASK).
The following monitored variables are necessary for further analysis:
resource depletion per task (TASK monitoring program)
The monitoring program provides:CPU time requirement
Number of SVC calls
the number of I/O operations and I/O times, broken down according to the volumes defined in the monitoring program
the duration of wait states
Paging activities
the message lengths per terminal I/O.
An additional user task measurement with subsequent analysis by SM2-PA (program analyzer) is recommended for analyzing the CPU requirements of individual tasks in more detail using program counter statistics and SVC statistics.
SVC statistics (SVC monitoring program)
The monitoring program provides the frequency of each SVC call classified by processor states TU and TPR.DAB hit rate (DAB monitoring program)
The monitoring program provides the number of accesses per specified DAB subarea.wait times in front of individual devices (SERVICETIME monitoring program)
ISAM pool hit rate (when NK-ISAM is used; ISAM monitoring program) The monitoring program provides the number of accesses per specified SAM pool.
number of access operations to particular files (FILE monitoring program)
UTM statistics (UTM monitoring program)
The monitoring outputs data for each application, including the dwell time in UTM, the dwell time broken down according to database share, TAC-class wait time and time for distributed processing. In addition, the program provides average consumption values for CPU and I/O operations in UTM and in the database system (only SESAM/SQL and UDS), as well as the number of database calls for each dialog step or asynchronous process.UTM values are output under the following conditions:
the UTM-SM2 subsystem must have been started
UTM-SM2 is started automatically when the operand MAX SM2=ON is specified for the KDCDEF run.
UTM-SM2 is started subsequently via the UTM administration interface with KDCAPPL and SM2=ON, together with the readiness to pass values.the SM2 monitoring program UTM must have been activated
the UTM application must have been set to supply data to SM2.
The passing of values is controlled by means of the MAX SM2=ON/OFF/NO operand in the KDCDEF run as follows:if SM2=OFF (default value) is set, the passing of values for each application can be activated at a later point in time via the UTM administration interface with KDCAPPL and SM2=ON;
if SM2=ON is set, no further administration is required;
if SM2=NO is set, the passing of values is generally prevented and cannot be activated at a later point in time.
database-specific consumption values (I/O operations, CPU time) are output only under the following conditions:
BS2000 accounting is activated;
The UTM accounting record UTMA is enabled (with
/MODIFY-ACCOUNTING-PARAMETERS, ADD-RECORD-TYPE=UTMA).UTM accounting is activated (KDCAPPL, parameter ACC=ON);
in SESAM/SQL, the recording of statistics is activated with the statement ACC,TP=ON,CPU.
Further information is provided in the manual “openUTM - Concepts and Functions” manual [19 (Related publications)] and in the manual “Using openUTM Applications under BS2000” [22 (Related publications)].
Database statistics
The SESAM-SQL and UDS-SQL monitoring programs supply monitored data about the SESAM/SQL and UDS/SQL database systems.Prerequisite in SESAM/SQL:
To transfer statistical data from SESAM/SQL to openSM2, start SESMON in batch mode:
/START-SESAM-PERF-MONITOR//SET-MONITOR-OPTIONS ...,OUTPUT=*SM2When OUTPUT=*SM2, only one DBH can be specified for each SESMON instance. A new SESMON instance must be started for each further DBH for which data is to be output.
The interval at which SESMON transfers the data to openSM2 is automatically set to approx. 30% of the SM2 monitoring cycle. It cannot be set manually.
Prerequisite in UDS/SQL:
Transfer of the statistical data from UDS/SQL toopenSM2 is initiated either when the UDS monitor is started using the MEDIUM=S,n statement or, during ongoing operation, with the /INFORM-PROGRAM MSG='ADD MEDIUM=S'n' command. It can be terminated again using the /INFORM-PROGRAM MSG='FINISH MEDIUM=S' command.
nis used to define the cycle in seconds (5 ≤ n ≤ 999) in which the UDS monitor transfers the data to SM2. It should be set to considerably less than the monitoring cycle set in SM2 so that data can be transferred several times in an SM2 monitoring cycle.The monitored values are supplied asynchronously to openSM2 from the database systems and apply for cycles defined for one or more database systems which do not need to match the SM2 cycle exactly. Here there can be differences in both the length of the cycles and in the temporal displacements between the database system and the SM2 cycles.
The length of the database system cycle or cycles is used to standardize the monitored values to one second. The values are therefore exact, but they only match the SM2 cycle to a certain degree.
Information on TCP/IP connections (TCP-IP monitoring program)
The IP and PORT number is supplied for each connection together with the number of transport service data units transferred.Number of accesses and data transferred of network devices
(monitoring program SAMPLING-DEVICE)Communication in the computer network (MSCF monitoring program) Data is supplied on the communication of the local computer with other computers.
Data on basic functions in the HIPLEX network (NSM monitoring program)
Statistics on lock requests from TU, TPR and NSM
(DLM monitoring program)POSIX statistics (POSIX monitoring program)
Supplies measurements on the use of various POSIX functions.