Limited number of workspaces vs workspace for department or line of business ?

[Avvar2022]

We are just getting started with databricks currently we have 1 workspace for each environment (DEV, QA and PRD). we have started with 1 workspace but there is already getting flooded with new workspace requests?

is there any check list/criteria for creating separate cluster vs separate workspace ?

Pros and cons of setting limited number of workspaces vs workspace for department or line of business ?

is there benefit having limited workspaces ?

is there any overhead of having workspace for each use case or department ?

Thank you in advance !

[Mounika_Tarigop]

I believe DEV, QA and PRD are the right segregate . We may need this because this will be easy to categorize the production and QA workload - based upon the amount of data cluster processes (Meaning more DBU) which we can restrict by the company workload and testing facility. Also, this will helps to differentiate the user level access (for example - engineers who does not need any access to prod environment) - and the list goes on.

so for above i think it's clear the criteria of having separate workspace, being said that - we can differentiate the cluster choice based on the department/ number of notebook / jobs cluster is handling and the nature of the workflow you are running inside a same workspace. (cluster segregate basically the nature of your workload )