Publish all the layers from several different SDEs - Best practices?

[AndrasNagySCC]

Currently we have a SQL database. We have different schemas on that DB, separating the different datasets (e.g. HIGHWAYSDATA, AUTOCADDATA, etc). Our userbase are using DB connections to access the data, and their access is controlled via Active Directory. 

We are moving away from Desktop GIS for the majority of the userbase, and moving them to WebGIS, using our Portal as their main tool to query the data we store. 

I am looking into the best way to push all the data we currently have on the SQL database into portal. 

Current idea:
1) Publish each layer 1 by 1 as a Map Image Web layer. 

2) Order the published layers into groups, just as they are sitting in the schemas currently. 

3) Give access to users only to their relevant groups. 

Problem with this idea: each layer needs to be loaded into a Map, then that Map needs to be renamed to the layer name, so when people are looking up the data in Portal, it won't look like Map/XY_Layer. The result of this would be XY_Layer/XY_Layer, which is still not optimal. 

I just started tinkering with ArcPy, making my first script yesterday overwriting an existing Weblayer, but I aim to automate this process so I don't have to manually do it approx. 200 times on the Test then 200 times on the Prod environment. 

Any advice or help would be very much appreciated! 

[JoëlHempenius3]

I think you want to use the Bulk Publishing feature:

Publish layers in bulk from a user-managed data store—Portal for ArcGIS | Documentation for ArcGIS E...

I don't think you need to add all the layers to individual maps. Users can do that themselves when they are making the maps or your users can load these services into their ArcGIS Pro Maps using the Add Data button and browsing the Portal

[Laura]

I grouped my layers in Pro and shared them that way: water valves, MH, main, cutoffs, etc all were published together and named water group in my portal. I did the same for sanitary, storm, parks. This seemed better than 1 by 1, which is what we did previously.