This message was deleted.

Hi Herpiko - (if I understand your question correctly) - GrowthBook does not cache any results of an assignment - instead, as you have figured out, we use deterministic hashing to make sure that if the user id (or whatever hashing attribute you’re using) is the same, the user will get the same variation. This means that if you spawn 5 replicas of a backend, the user will get the same variation for all/any of them.

@fresh-football-47124 Thank you for your response! It seems that you understand my question. All of this deterministic hashing are happen locally without the Growthbook server, right? The main purpose of the Growthbook dashboard is to "configure" and provide the values/experiment config via API. The distributing works themselves are being handled by the client locally. Please correct me if I am wrong.

correct, it’s done locally

@careful-table-10580 Thank you!