For the Event Store, what were the requirements you were looking for?
Functionally, Kafka just allows you to replay logs from a specific offset. Is this enough for an event store? Wouldn’t you also want to look up all the events as well?
From what I understand, Kafka here just decouples your write availability with the read availability.
In one of the comments, you’ve mentioned that the retention period is in years, and you are not compacting the Kafka logs. Would it make sense to also persist all the offsets for a given entity in another system? This will help you replay events for an entity by querying up the relevant offsets. I am not sure if this is feasible though.