Are we misuing the ECS "event" fields?

@cmoussa1 @cdefrates I wonder if we are misusing the "event.*" fields in the HPC Job Scheduler Event Schema. Granted, the definition of an "event" is a bit nebulous, but in my mind events are more of a discrete-moment-in-time sort of thing. And events have "duration", and start/end times in ECS, so that starts to make them look like something we can use/abuse for our purposes. But it may not be correct to model long-running activities as "events".

For instance, I would say that the HPC Job Scheduler Event Schema is really associated with the job completion event. And we should keep in mind that there are lots of other events that we will be capturing for a job during its lifetime: job submission, schedule, prologue, launch, epilogue, etc.

Maybe the start/stop/duration of the job should be moved under the "job." namespace? That would free up the event.* fields to measure things associate with just the job completion event.

Start by reviewing the HPC Job Scheduler Event Schema and the existing event.* and job.* fields referenced in the issue. Compare those fields with the job lifecycle events listed—submission, scheduling, prologue, launch, epilogue, and completion. Done means reaching and documenting a decision on where long-running job timing belongs.