I am facing a performance issue and memory limit on my aggregation pipeline. My goal is to extract documents from "tiles" collection, each tile storing a time series f
