Yeah, I get that.
What I’m saying here is, because multiple processes try to insert the job at about the same time (single digit ms appart) apparently, uniqueness isn’t guaranteed at insert time either?
I ran an experiment of spawning 1000 parallel tasks, each inserting the same job with the same arguments, and while not 1000 get inserted, it’s definitely in the range of several tens of them.
I guess i understand why that’s the case, but I’m wondering if I’m missing something.
The chain option might at least allow me they don’t execute at the same time, even if they are duplicate inserts, right?
