Many Scheduler administrators have been required to produce a report on different statistics about the processing done by Scheduler. Scheduler keeps track of how many times certain internal tasks occur, and how much time is spent doing these tasks and reporting is now available.
Release: 11.0
Component: Scheduler Job Management MVS
The most common request for statistics is the number of jobs completed, amended or failed daily that has been processed by Scheduler. The JCL below will accomplish this request, and includes data reported at 8 hour shift increment.
//STEP1 EXEC CAJUTIL0
//SYSIN DD *
REPORT STATS I=480,FD=-2,SYSID=SYSA,ID='MNTR JOB COMP',FT=0800
REPORT STATS I=480,FD=-2,SYSID=SYSA,ID='MNTR JOB ABND',FT=0800
REPORT STATS I=480,FD=-2,SYSID=SYSA,ID='MNTR JOB FAIL',FT=0800
//
I = reporting interval in minutes
FT = start time of the first day of reporting
FD = negative number of days which includes today's data or a Gregorian date with either a 2 or 4 digit year
SYSID = change to the SYSID of the MCPUSUB for these statistics, because they only occur on the MCPUSUB. ENF events, on the other hand, happen on every system and therefore each system has its own stats.
Here is the result of the report.
REPORT STATS I=480,FD=-2,SYSID=SYSA,ID='MNTR JOB COMP',FT=0800
STATISTIC NAME=MNTR JOB COMP FOR DATE=09/18/2006 SYSID=SYSA
0800-1559 12 1600-2359 2
STATISTIC NAME=MNTR JOB COMP FOR DATE=09/19/2006 SYSID=SYSA
0000-0759 7
0800-1559 0
1600-2359 3
STATISTIC NAME=MNTR JOB COMP FOR DATE=09/20/2006 SYSID=SYSA
0000-0759 9
0800-1500 24
REPORT STATS I=480,FD=-2,SYSID=SYSA,ID='MNTR JOB ABND',FT=0800
CACU060I NO STATISTICS FOR 09/18/2006
CACU060I NO STATISTICS FOR 09/19/2006
STATISTIC NAME=MNTR JOB ABND FOR DATE=09/20/2006 SYSID=SYSA
0000-0759 0
0800-1500 1
REPORT STATS I=480,FD=-2,SYSID=SYSA,ID='MNTR JOB FAIL',FT=0800
STATISTIC NAME=MNTR JOB FAIL FOR DATE=09/18/2006 SYSID=SYSA
0800-1559 1
1600-2359 1
CACU060I NO STATISTICS FOR 09/19/2006
CACU060I NO STATISTICS FOR 09/20/2006
Using the same report and altering I=60 will give the hourly data that can be used for trending on peaks of processing for daily workload. Changing the I=1440 and FD=-365 produces a daily report for monitoring the last years trend in changes of volume of completed, amended and failed jobs processed by Scheduler.
Scheduler spends most of the time waiting for events to occur which should not incur significant overhead. There are some exceptions which can impact performance such as Autoscan. Reviewing the number of backlogged schedules daily will determine if Autoscan is occurring during a slow period of production processing. Additionally, time needed to perform the Autoscan is also available.
Reporting on ENF DSCLOSE, ENF DSCLOSE USE, ENF JOBINIT, ENF JOBINIT USE or other ENF events will show the events used versus received. If the results have a very low percent of use, review the SCREEN or SELECT statements in ENF. Isolating the events needed for Scheduler will reduce unnecessary events to be processed by ENF.
Using LIBTYPE=RDRQ can add to overhead. Analyze the report produced by MNTR RDRQ SEARCH, MNTR RDRQ NTFND and MNTR RDRQ TIME with I=60. Specify an early start time for the job as close to the submission of the job as possible.
Daily statistics are kept for 2 years if using the default option and can be altered in your options to save for 5 years. Report Statistics covers a variety of data regarding e-mail, cross platform, JCL, library type submission and more.