Group data sources

Data sources should be grouped so that multiple connections to the same data source can be identified as similar.

This would be useful if e.g. two connections to the same data source are created with different external API keys with different permissions