MicrosoftHi Mark Kromer
Any news on that 'WIP' connection parameterisation feature you mentioned? I noticed Fabric has had a 'little' update recently in this regard where you can now dynamically choose the name of a connection. I almost thought they'd given us a possible workaround instead of giving us the parameterised the server name feature (which is really what is needed, like you could do with old ADF linked services), but then I realised it only works when combined with a limited choice of only Fabric items such as a lakehouse or warehouse as the 'type'...
Ah this was so close, if only it had also allowed SQL server and other possible types that are common on-prem, that at least would have been a kind of workaround I could live with 
I've had to recommend that a number of my clients use Azure databricks in the meantime combined with vanilla ADF and a metadata driven architecture - with a possible future migration to Fabric. Although I have introduced a small metadata driven Fabric solution for a few clients, but only to those with just the one (or very few) on-prem source servers. For some of my clients who are reaching near 100 servers its just not practical, unless we build them some sort of ADF/Fabric hybrid architecture. However, most of them seem happy to use just Azure databricks with ADF in that regard, as its more mature...
So I guess we're still waiting for the 'server name' parameterisation feature... unless the devs can expand upon the dynamic 'connection name' feature that was introduced?
P.S. I've voted for this feature on the 'ideas' site https://ideas.fabric.microsoft.com/ideas/idea/?ideaid=cd5ccbf7-391c-ee11-a81c-6045bd7e3068, but it seems it needs more votes
