Design-time information based on run-time artifacts in transient cloud-based distributed computing clusters

This invention describes a system for understanding how temporary cloud computing resources are used. It works by collecting data about what these temporary clusters do while they're running, even after they've been shut down. This collected data is then analyzed to figure out patterns, dependencies, or inefficiencies in the tasks that were performed, specifically for a single workflow. This understanding can then be used to improve how that workflow is designed or how future computing resources are allocated for it.