Question for the gurus: examples for using spicedb, and how I'm using it for testing use a single token related to accessing the system. 'serve-testing' lets you provide different tokens for 'clean slate' behaviour. But, what about a case where I would like to have different access tokens for different services who are publishing relationships?

Are these services writing relationships to the same permission system? If yes, what's your use case for having access tokens for each service?

They will be, but they are responsible for very distinct subsets of resources; they may create a relationship to one or two other definitions in the schema, but rarely. What I really want is just to have some sort of 'identity' for the services doing the operations. Be great to be able to track which is doing what and when.

We don't have a solution for that exact problem right now. We do implement something called Fine Grained Access Management (FGAM) in our SpiceDB Dedicated offering. FGAM allows you to provision tokens with varying sets of permissions. Though, FGAM is not available for open-source users. You can read more about FGAM here: https://authzed.com/docs/spicedb-dedicated/fgam