It looks like, in order to run topology or a network on a feature class(es), they need to be in the same dataset and obviously have the same parameters (spatial reference etc).
To run these, the feature classes either need to be versioned, or non-versioned *without* archive enabled.
But you can't publish a sync-enabled feature service from a non-versioned data source unless you have archive enabled.
Sooooo, if we don't want the data versioned, is there any other workflow for running topology or network analysis on a feature class that is non-versioned and archive enabled?
My only thoughts were a replica database somehow but I'm not too sure what that workflow would look like.
Would be interested to hear from anyone who might have attempted to do something similar and any potential solutions.