sql server – Structuring Database projects for use in Azure DevOps

We have a number of SQL Server databases that are not currently under source control. We are looking to use Azure Repos Git to address this.

Some of our databases contain cross database references such as querying another database View or table. I am therefore curious on what the best practice is for structuring our solutions and/or code to give best flexibility for use in azure CI/CD?

For example – should we use 1 Visual studio solution per database project? or One solution to hold all our database projects? Just looking to get a feel for the pros and cons before we invest in one particular method over another