As a Dataiku admin, I have set up some connections to variety of databases using functional / generic accounts. In the connection 'Security settings', I have added 'usable by groups' the group that I would let them to use those connections.
We noticed that developers in the group can only use the connections to create datasets, but they can't view those connections under menu of Administration -> Connections,
I then tried to add the group under the 'Details readable by' under the 'Selected groups' for a connection. It still not show up for them, even though the description sounds like it is for this purpose!
Any suggestions? Our developers would like to view all the possible connections that admin has set up to use. Thanks!
there is no setting to grant access to the connection pages in Administration > Connections to non-admin users. The "details readable by" setting only controls which connections can be passed to recipes for direct access by the recipe process.
The notion of "connections available to the user" is governed by the "freely usable by" setting. When users create datasets in their flow, the list of connections they're presented with is filtered according to this setting.
So in short, non admin users can't see in the UI the setup of the connections they're allowed to use, by design.
Hi, Thank you for quick response! Could you elaborate a little more on "which connections can be passed to recipes for direct access by the recipe process", maybe through a quick example?
We have set "details readable by" to 'Nobody' for all connections currently. People are able to use the connections to query and create datasets in a flow, or query inside python recipes using code example below, or in webapps. What is the real purpose of that setting seems still no clear? Thanks!
executor = dataiku.core.sql.SQLExecutor2(connection=connection)
df = executor.query_to_df(query)
the purpose of the "details readable by" settings is explained in https://doc.dataiku.com/dss/latest/security/connections.html#reading-details-of-a-connection .
Its main purpose is to let recipes running on Spark use fast-paths to access the data.