Discoverability and request access for all item types

[niels-janos_sim]

Currently there is a feature for semantic models to publish them in the OneLake Datahub for people to be able to discover them and request access. Similarly it would be great to also be able to promote Lakehouses (and Data Warehouses, Notebooks, environments,...) to be discoverable by user who do not yet have access. This would enable the Datahub to become a bit like a marketplace where I can promote my Data to potential consumers (or share my methodology in the case of Notebooks, environments)

[PaulAiimi]

An interesting requirement has come out from a client I am engaged with where they want users to discover ALL content that is out there within the tenant. Where there is a de-centralised approach to analytics (Self-Service), we want to make sure we break down barriers and make sure duplication is kept to a minimum. From a development perspective, this is managed. But from an end user perspective, they can only search for what they have access to in the first place. What would be great is to show case; - What Reports/Dashboards/Apps are out there - Whether the user has access to it - Who is the Information Owner so that they can request access If they have access to any then great, otherwise show the information about the App and the owner. To mitigate any potential data breaches, the App could be published with a toggle to suggest whether people could see what is in there (from a meta data perspective) or not if the content is highly sensitive within the 'Publish to App' option. For end users, maybe an extension to the current App Discovery route (Get Apps).

[David_Whitney]

In dataset settings there's a section for "Endorsement and discovery" where in addition to making a dataset "Promoted" you can also make it "Discoverable." Dataflows currently only allow "Promotion" and not "Discovery". After all, the Microsoft documentation on Dataset Discoverability says, "Without discoverability, the full value of endorsement, that is, directing users to quality data, is not fully realized."

[nschachter]

Looking into this idea. More input and votes from the community would be helpful.

[Erik_Eld]

Fully agree with previous writers. We are making Warehouses available as "OneLake Data Products" from which business can build their own Power BI models/reports, combine to create new data products or consume in other ways. As warehouses cannot be made Discoverable, we must maintain a catalog using other means than the OneLake catalog. Some might say that we should use semantic models instead for this. Those have apparently not tried it on a F64 capacity in larger scale with Direct Query performance issues, Direct Lake limitations etc. For us, it's given that all items in Fabric should have the option to be made Discoverable and Endorsed. As it is now, the OneLake catalog is not really a catalog, as it mainly shows what you already have access to.