Is there a way to run notebooks concurrently in same session?

[darshan]

tried using-

dbutils.notebook.run(notebook.path, notebook.timeout, notebook.parameters)

but it takes 20 seconds to start new session. %run uses same session but cannot figure out how to use it to run notebooks concurrently.

[Hubert-Dudek · Accepted solution]

dbutils.notebook.run starts a new job, that's why it takes this time and test yo can start multiple concurrently using ThreadPool or other async libraries. Probably with the better server, it could be 10 seconds, not 20.

%RUN executes other notebooks like it would be code of notebook which we run (like include in some languages).

Maybe you could plan a different architecture way (but it depends on the use case) as this way is instead for no-time critical projects (analytics). Stream and delta live tables should be used instead if it is a project when that 20s delay is too much.

[craig_lukasik]

Another approach is to use the Jobs API and leverage a notebook_task with an existing cluster (existing_cluster_id).

[Prabakar]

You can use jobs workflow. It helps you to run the notebooks in parallel as well as sequential.

[darshan]

Is there a way to send data from one notebook to other using workflow?

[bb_huey99]

This is a great question. I’ve been struggling with opening multiple browser sessions to open more than one notebook at a time.

[sbo]

dbutils.notebook.run will run multiple notebooks in the same session!

[darshan]

But it creates new jobs which makes process slow by 20 seconds

[Kaniz]

Hi @darshan doshi​, We haven’t heard from you on the last responses, and I was checking back to see if those suggestions helped you. Or else, If you have any solution, please share it with the community as it can be helpful to others.

[darshan]

Hello, yes responses were helpful, so I tried workflow but stuck with sending data across notebooks using workflow.

Also, trying delta live tables. What I would like to know is is there threshold about how many records delta live tables can hold at a time?

[darshan]

Hello, one thing I am not able to understand is, while developing low latency application where even 5 seconds cannot be tolerated, how to create jobs? Because each job takes 20 seconds to init.

[zulazuardi]

You can use python concurrent run and execute the dbutils.notebooks.run​

[gfree76]

There is a databricks sample notebook that shows how to use scala futures to run notebooks in parallel. We use this extensively, so I can promise it works 🙂 Look for the section Run Notebooks Concurrently https://docs.databricks.com/notebooks/notebook-workflows.html

[darshan]

This does help, but can you help me understand why job run time takes more time than actual run time of notebook? And is there a way to reduce the time?

[rudesingh56]

I’ve been struggling with opening multiple browser sessions to open more than one notebook at a time.