mma-retry

To retry failed tasks from a dispatch, provide the original taskId and an array of zero-based taskIndices to specify which tasks to re-execute, isolating failures while preserving original diagnostics.

Task-level retry orchestration is the process of re-running selected failed tasks by index rather than re-executing an entire dispatch. You need it to isolate and fix failures efficiently without losing original configuration data.

To identify tasks to re-run, provide a zero-based taskIndices array alongside the original taskId. This targets specific failed or completed tasks for re-execution while preserving the original dispatch configuration.

Retrying selected tasks preserves original task diagnostics for comparison and auditing. The retry mechanism targets specific tasks by index while maintaining the initial configuration and returning a fresh taskId.

To monitor a partial dispatch retry, submit the original taskId and taskIndices to the retry endpoint. You receive a new taskId for the retry run to track progress while keeping original diagnostics intact.

Your task retry returns a new taskId to monitor the fresh retry run separately while keeping original dispatch diagnostics preserved. This isolates the re-executed tasks from the initial completed or failed dispatch.