A job has a run_window and date conditions that have have it scheduled to run at the same minute as the close of the run_window. Normally, that last scheduled run at the close of the run_window works fine as it is still within the run_window. However, if there is any delay processing the STARTJOB event for that last scheduled start at the close of the run_window, the job will not run. The scheduler considers it outside the run_window instead of realizing that it was supposed to start within the run_window.
Release : 11.3.6/12.0
Component : CA Workload Automation AE (AutoSys)
When the scheduler checks the run_window for a job that it is starting, it is not going to look back at the original event time for the STARTJOB event. It is only going to look at current time and only start the job if current time is within the run_window range. There are use cases where looking back at the original event time for the STARTJOB event would not be desired and allow a job to run far outside its defined run_window. For this reason, it is always best practice to define the close of a run_window so that it doesn't match a possible scheduled start time for the job. Set it to 1 or 2 minutes after the last possible scheduled start to allow for any volume related Scheduler latency.