I have a special user that has observed a delay in triggering his jobs sometimes up to 10 seconds, and it has at times caused the application to miss their SLA commitment of completing their reports on time. I have used CA7 far longer than I want to admit but I have never bothered to validate the triggering timeframe up to the second, rather I have tracked the whole
workflow to ensure it meets the OLA or SLA for that perticular business need.
BTW - this user thinks it should be milasecond triggering event. This CA7 instance is set up on one machine (where this workflow exists) that goes directly to the internal reader, and also has a submit dataset that will pass jobs to another LPAR that is set up under a differnt time.