Feature request: add Bucket resource

[cortopy]

This is just a feature request for NACK to handle Key/Value buckets in addition to Consumer and Stream.

[Deimos2]

This should be a feature, but you can also create a k/v bucket from a Stream as mentioned in https://github.com/nats-io/nats-architecture-and-design/blob/main/adr/ADR-8.md#buckets

[dwooldridge-ctoken]

Hi, it would be really nice to have a KV Bucket resource to work with. Do you think this is something that may be implemented?

[Jarema]

Yes. It is.
However, first we want to finish the rework of nack internals.

[dozsa]

Any update on when K/V bucket setup would be supported? +1 on the usefulness of such a feature. Thanks.

[SimonSchmid]

+1

[praveenperera]

Here is an example of how to create a bucket

---
apiVersion: jetstream.nats.io/v1beta2
kind: Stream
metadata:
  name: task-store
  namespace: nats
spec:
  name: KV_v1-task-store-borsh
  account: nack
  subjects: ["$KV.v1-task-store-borsh.>"]
  maxMsgsPerSubject: 1
  storage: file
  maxAge: 200h
  retention: limits
  duplicateWindow: 1m
  replicas: 3

[dozsa]

That is an interesting workaround, to create the underlying stream directly. But ideally the implementation details of the bucket should be hidden under a proper bucket CRD abstraction. One problem with creating the stream directly is that if the stream details change in the KV logic in the drivers you could end up with strange behaviors.