Generate performance reports based on *HSTFILE, and also have the possibility of automate report generation if possible.

Hi

I agree, having a scheduler to produce collections of these performance reports would be a great idea.

I wanted to check that you knew you can write your own, all PDi graphs including the Graph History ones have an option to show you the SQL that was used to generate the graph. You can take this SQL and use your own job to generate it. An example at the bottom of this comment.

Hope this helps. #StaySafe
Steve B
IBM Champion, CEAC Member and Technical Director of i-UG.co.uk

/** Historic Composite chart: CPU Utilization (Overview) */
SELECT QSYS.DATETIME AS DATETIME,
SYSSPTU,
SYSCTA,
COLPCTCPU,
SYSPTU,
COLPCTSCPU,
SPURR_PURR_RATIO,
CASE
WHEN QSYS.NUM_DATA_POINTS > 1
THEN 'NO'
ELSE 'YES'
END AS "IS_SINGLE_POINT",
CASE
WHEN QSYSDET.DATETIME2 IS NULL
THEN 'NO'
ELSE 'YES'
END AS "HAS_DETAILED_DATA"
FROM (
SELECT MIN(DATETIME) DATETIME,
AVG(SYSSPTU) AS SYSSPTU,
AVG(SYSCTA) AS SYSCTA,
AVG(SYSPTU / DOUBLE (SYSCTA) * 100) AS COLPCTCPU,
AVG(SYSPTU) AS SYSPTU,
AVG(SYSSPTU / DOUBLE (SYSCTA) * 100) AS COLPCTSCPU,
AVG(DOUBLE (SYSSPTU / SYSPTU)) AS SPURR_PURR_RATIO,
COUNT(*) AS NUM_DATA_POINTS
FROM QPFRHIST.qapmhmsyst
WHERE DATETIME >= '2020-04-24 00:00:00'
AND DATETIME <= '2020-05-24 00:00:00'
GROUP BY YEAR(DATETIME),MONTH(DATETIME),DAY(DATETIME),HOUR(DATETIME) / 6
) QSYS
LEFT JOIN (
SELECT DISTINCT j.DATETIME AS DATETIME2
FROM QPFRHIST.QAPMHDJOBM j,
QPFRHIST.QAPMHMSYST s
WHERE s.DATETIME = j.DATETIME
) QSYSDET ON QSYS.DATETIME = QSYSDET.DATETIME2
ORDER BY DATETIME