EDIT: I saw a few demos and I think I can answer myself here: ECK makes it so much easier to maintain a single multi-node ELK cluster. Rolling updates to new Elastic stack versions, auto scaling, etc.
I really think we should not only support ECK, but use it also for daily development and integration tests.
@Romain to be honest, me too. Quite a few customers have requested for ECK interop.
There is one thing I didn’t fully understand though (and maybe you can answer).
Typical ROR Enterprise sales pitch
ROR Enterprise comes with pretty good segregation capabilities both on data access and Kibana configuration, without the need to instantiate multiple clusters for multiple tenancies.
And this introduces great savings in terms of compute resources: everyone can work with a large, capable ELK cluster instead of having their own tiny cluster.
Interest in ECK?
But now I see customers (even Enterprise customers) asking for ECK, which is basically a GUI for automatic creation of multiple ELK clusters in Kubernetes.
Poll: what’s the logic behind ROR users interest in ECK?
Saving in compute resources introduced by ROR Enterprise are less than the license price?
There’s a stricter requirement in physically separating data from different tenants?
You already use Kubernetes, and ECK is seen as a clean way to install the single, ROR-powered multi-tenant cluster?
Something else (comment please!)