Databricks Jobs and CICD

[labromb]

Hi,

We currently leverage Azure DevOps to source control our notebooks and use CICD to publish the notebooks to different environments and this works very well.

We do not have the same functionality available for Databricks jobs (the ability to source control or deploy through CICD). Could we call the Databricks Jobs create API to promote the job definitions through environments or is there a better way of doing this?

Look forward to hearing from you

[labromb]

Hi @Kaniz Fatma​ interesting new feature for sure but not sure how this helps me with the challenge. As an aside, there probably also needs to be something similar for Databricks tables as the definition of the tables is essentially code.

Echoing @Anders Bergmål-Manneråk​ comments from yesterday....

[bergmaal]

We're also missing this feature to have easy version control and CI/CD for Jobs/Workflows. The answer from @Kaniz Fatma​ does not cover the question as I see it.

Right now the only options we see is using Terraform, or the Jobs API 2.1 as superficially described in this post. But they both require some sort of custom setup. What we really need is a more integrated experience like we have for notebooks using Databricks repos.

It would be great to get some updates from Databricks Workflows team on this. Is this a feature on the roadmap? If not, it would be awesome to get some official advise on handling source control and CI/CD for Workflows/Jobs.

[saadansari-db]

Hi @Brian Labrom​ and @Anders Bergmål-Manneråk​ - my team is looking at creating a set of features to provide simpler source control and CI/CD. Can you please send me an email at saad dot ansari at databricks dot com?

I would love to get more feedback incorporated into our approach here.

[bergmaal]

@Saad Ansari​ popped you an email now.

[labromb]

@Saad Ansari​ ​ have done the same

[JRT5933]

My team is currently looking at establishing REPO(s) for source control to start. I know I've seen some documentation for when a MERGE is completed to auto update the main branch in DBX remote repo. Does annyone have a template and/or best practices here?