e.g. if two flows running on different instances in my cluster happen to call the same task, it's not a big deal if they both run. if its 10,000 of the same call across a couple minutes, it would be awesome if i those results were stored in a distributed cache so that each task can just check said distributed cache