Assign + SR

[JoranVanBelle]

While using this library some questions came up to me.
(1) Is there already a way to fetch the schema registry using HasKafka, or is there any other way to do it, so avro messages, as well as json can be consumed(and published).
(2) I find it a bit weird to use the assign of a consumer when you already have to subscribe to a topic when you want to create a consumer. Is there a way to create a consumer without having to subscribe up front?

[AlexeyRaga]

1. Perhaps you are looking for https://github.com/haskell-works/hw-kafka-avro @JoranVanBelle ?

2. I don't understand... You don't need to use assign, if you just want to subscribe? It only make sense to use in a callback if you have some logic to perform on assignment..
   Or do you want to assign partitions without subscribing? But then vow would you know which ones? Say, the job auto-scales, how do you determine what partitions to assign for which instance? You end up re-implementing rebalancing protocol?

[JoranVanBelle]

Ah okayy, there is another library dealing with avro, thanks for sharing @AlexeyRaga!

You have use cases where you don't want to rely on the rebalance. If you use some kind of monitoring tool like conduktor, you can't subscribe to a topic. Because therefor you need a consumer group which won't work in a production environnement as you cannot read from any topic if you're not giving it the right permissions. If you want to read from a topic (and not trigger a rebalance), you have to consume without consumergroup and use assign to read from the partitions.
So I was wondering if there is already a way to create a consumer without a consumergroup/Subscription. Because I think assign is the most powerful to use if you don't rely on a consumer group as you then don't have any rebalance.