Job or Task backtracking - Focusing on Task Granularity for Backtracking, Not Workflow Granularity

[gwdgithubnom]

To ensure fault tolerance, backtracking should be performed at the Task level, not only at the Workflow level in DS UI platform. This allows for the recovery of individual failed tasks without requiring the restart of the entire workflow.

Due to business needs, complex workflows often have multiple task nodes. Currently, if a specific task fails or produces incorrect results, we need to re-run the entire workflow, which is inefficient. Therefore, if we need to backtrack only a portion of the tasks instead of the entire workflow, we have to develop custom interfaces to invoke the platform work. Are there any plans for Dolphinscheduler to implement this functionality in the future?