You have several ways to help you ensure compliance, as many of our users have:
First, recognize that Pub/Sub makes every effort to store data closest to the source of the publication. Once stored, the data is not moved to another storage location without explicit action by the user. Practically this means that it’s is highly unlikely for your data to be stored far from the publisher client.
Second, Pub/Sub provides monitoring on data storage location (size of backlog by region, for example). You can, therefore, monitor for potential violations and take action by discarding the backlog using Pub/Sub's seek's functionality or ensuring that it is quickly processed.
Third, you can limit your risk by reducing message retention duration.
Finally, we are working on a more direct, explicit set of controls. We expect to share detail on this in the coming months.
Kir