This is important for ETL/ELT workloads:
You may want certain jobs to run in parallel. This is possible when your jobs are not dependent on each other. (It’s also more time efficient)
Other steps will be dependent on previous steps first completing successfully.
Multiple pipelines can be defined. This is useful eg: one of your pipelines could be for a production ELT process whilst another could be for exploratory ad-hoc purposes.
Each pipeline is a separate sequence of steps. Each pipeline has it’s own schedule.
You can schedule your pipeline to run up to a maximum of every 5 minutes. (This is subject to the pipeline being able to completely execute in this time-frame)
Example: If a pipeline takes 1 hour to complete, it will not be re-executed every 30 minutes. If you instruct your pipeline to run every 30 minutes, it will effectively skip every second run and only run once per hour (not twice per hour)