@tanepiper For any sufficiently complex cloud service that is built to scale, there are two possible states for a representative to be in when they tell you that data can be deleted from the service: 1) incorrect, 2) lying.
The chances that there is any one person in the organization who can exhaustively list all of the places the data goes is as close to zero as makes no odds.
@tanepiper such services tend to be Conway’s Law-shaped. Teams are a given size, and have domain over a slice of the data (as a microservice) and its relation to the other data (as a protocol). The other parts of the lifecycle are hard enough, and there’s a good chance deletion wasn’t originally part of the design.
Retrofitting deletion in a distributed system requires some additional cross-microservice eventing and messaging, which needs to be adopted everywhere.