Problem:
AIDP pipelines/jobs often participate in a broader data flow. Today, orchestration becomes unnecessarily complicated when Pipeline B should run only after Pipeline A has completed successfully.
A basic dependency capability should allow an administrator to define:
Pipeline A → Pipeline B → Pipeline C
with downstream execution occurring only after successful completion of the predecessor.
Suggested enhancement
Add a simple Depends On / Predecessor property to pipeline scheduling that supports:
- Run pipeline only after another pipeline succeeds
- Support multiple predecessors
- Do not trigger downstream pipeline if predecessor fails
- Display dependency status in monitoring
- Allow dependencies across custom augmentation / ingestion pipelines where applicable
- Optional configurable delay after predecessor completion
