I'm currently working on an application platform based on postgres and was very happy to hear that Azure will soon offer Citus scaling (or does already). During

I'm currently working on an application platform based on postgres and was very happy to hear that Azure will soon offer Citus scaling (or does already). During the development phase it doesn't make sense for me to spend a lot of money in order to have my database hosted in a sharding mode but in the long term I would like to use the Citus offer on Azure. Do you have any advice on what to consider in my database (table) design in order to avoid issues when migrating to Citus at one point in time? I only read that it makes sense to have one column (e.g. customerId) in each table so the sharding mechanism knows based on which information the data should be distributed among the nodes. In my case I guess such kind of key would be organizationId (= a group of users)