RE: Job run times occasionally seriously spike. How do I start to figure this out?

Hi Rob,

Being a rookie, I would do this:

I would not use Navigator for i. I am fed up with its erratic behaviour. (I
also encountered the same issue as you.) I would start with good ol' green
screen GO PERFORM. I would start to compare ERP* job's running times (using
the history log, filtering job start/end messages), to see when it started.
Then I would check performance data to see if there is a job, or LIC task
eating up performance during the jobs running time. Has the program ran by
those ERP* jobs changed? Is the running time in seconds showed by the
CPF1164 messages in the History log significantly shorter then the "real"
running time (end time -start time)? If yes, I would suspect some other
waits increase (DLYJOB statements in the application for instance.)

My two cents worth.

Regards,
Fred