Skip to main content

Synapse

New

Pre-start Spark Pool in Pipelines

Vote (2) Share
Joshua Rodgers's profile image

Joshua Rodgers on 09 Jul 2024 23:48:43

It would be great if we could start a spark session when a pipeline runs so it is already ready when notebook tasks kick off. Currently, there is a delay added to each notebook activity while the spark session starts up.


A common pattern below would greatly benefit from having the spark session start prior to the notebook activity starting.


Current:

1-Pipeline Starts

2-Copy Activity

3-Notebook Activity (1-3 minutes delay with sessions startup)


Proposed:

1-Pipeline Starts

2-Spark Session Starts

3-Copy Activity

4-Notebook Activity (no delay as session has already started)