MICS

 View Only
Expand all | Collapse all

MICS RESP statement in cccOPS member - accurate / used at all?

  • 1.  MICS RESP statement in cccOPS member - accurate / used at all?

    Posted May 21, 2015 11:23 AM

    The current MICS-distributed cccOPS members have a RESP statement that warrants review / dial-down to define more useful and accurate "response distribution limits" (to multiple sub-second intervals) than what's provided by CA.  Take a look at your cccOPS members and see what's defined.  And compare your unit database values to the distributed-version of cccOPS in sharedprefix.MICS.HOLD.PARMS(cccOPS).

     

    Of course this discussion does presume there is MICS reporting interest with exploiting MICS / transaction percentile analysis using summary-level MICS database reporting files, like CICCSU, CICCSY, and others for IDMS and IMS.  Don't you want to be able to inquire with CA MICS and address the question: "99 percent of all CICS application *** transactions completed with 0.5 seconds" yesterday, or on-average for the past 53 weeks or 6 months.....okay, I'll stop there.

     

    And so here's a tip with MICS CICS Analyzer to substitute the current response limit 1 for the appropriate SAS LABEL based on what's defined in MICS...after generating this report, then it may be time to reconsider the RESP settings, discuss them with end-users and decrease the limit-values so you can get more out of MICS (BTW, the code below can easily be applied to IMS or IDMS components with some minor changes to SAS variable names):

     

    //STEP1 EXEC MICSSHR?

    //SYSIN DD DATA,DLM=ZZ

    OPTIONS SOURCE SOURCE2 NOCENTER;

    %LET CYCRANGE = 01;

    * SORT/SUMMARIZE MICS CICCSU USING MICS SUMMARIZATION. ;

    *  SUMMARIZE TO ONE OBS PER CICS-REGION, PER DATE/HOUR. ;

    %LET BY = SYSID CSUAPPL INTVLHR;

    %*  code below captures last SAS variable from BY list. ;

    %LET BREAK = %SCAN(&BY,-1,%STR(&BY));

    DATA CSU;

      SET %MFILE(F=CSU&CYCRANGE,TS=DAYS);

      * *SBB*  Optional WHERE statement could go here. ;

      * YOUR CICCSU FILTER CODE GOES HERE, OTHERWISE USE CICCSY. ;

      * ;

      IF _N_=1 THEN DO;

        * DECLARE SAS MACRO VARIABLES FOR LATER USE IN LABEL STMT. ;

        CALL SYMPUT('RVAL1',PUT(CICRVAL1,6.2 -L) );

        CALL SYMPUT('RVAL2',PUT(CICRVAL2,6.2 -L) );

      END;

      * derive hour-interval for summary-reporting, convenient as one-column. ;

      ATTRIB INTVLHR FORMAT=DATETIME15. LABEL='INTVL-END HOUR';

      INTVLHR = INTNX('DTHOUR',ENDTS,0);

    RUN;

    PROC SORT DATA=CSU;

      BY &BY;

    RUN;

    %MACRO NEGATE;

      * MICS fffSUM macro invokes this code below to combine all CP-types ;

      *   just prior to OUTPUT. Remember to combine zIIP CP usage below,  ;

      *   starting with CICS/TS V5. ;

      ATTRIB AVGCPUTM FORMAT=10.3 LABEL='AVG CICS CPU P/TRAN (SEC)';

      AVGCPUTM = SUM(CSUCPUTM,CSUNZZTM) / SUM(CSUTRANS,CSUETRN);

    %MEND NEGATE;

    DATA CSU;

      %CSULEN(TS=MONTHS);

      SET CSU;

      %CSUSUM; /* NEGATE macro invoked just prior to OUTPUT */

    RUN;

    PROC PRINT UNIFORM NOOBS LABEL DATA=CSU;

      VAR &BY CSUTRANS CSUAVTTM AVGCPUTM CSUPCTR1-CSUPCTR2 CSUTDST1-CSUTDST2;

      FORMAT CSUTRANS COMMA9. CSUAVTTM 14.3 ;

      * OVERRIDE SAS LABEL BELOW - ADD MICS RESP DIST LIMIT. ;

      LABEL CSUPCTR1 = "PCT RESP W/IN &RVAL1 SEC";

      LABEL CSUPCTR2 = "PCT RESP W/IN &RVAL2 SEC";

     

      LABEL CSUTDST1 = "TOT RESP W/IN &RVAL1 SEC";

      LABEL CSUTDST2 = "TOT RESP W/IN &RVAL2 SEC";

    RUN;

    ZZ

     

     

    Another tip for ASG TMON/CICS (TCE) users (hey, now aren't you glad you read this far down in the post?) -- if you are at TCE 3.4 or later/higher, be sure that the TMON option is properly set to split-out your zIIP CPU consumption as separate reporting metrics, otherwise it gets lumped together with general-purpose CP consumption and cannot be segregated for transaction/application reporting interests.

     

     

    Scott Barry

    SBBWorks, Inc.

     



  • 2.  Re: MICS RESP statement in cccOPS member - accurate / used at all?

    Posted May 22, 2015 02:50 AM

    Good morning Scott,

     

    Thx a lot for sharing!

     

    Michal Koch