Ali Kanso

at Microsoft we use Azure Kubernetes Service, with enabled AKS container log collection to Log Analytics through [Azure Monitor](https://docs.microsoft.com/en-us/azure/azure-monitor/overview) for telemetry and logs. Using FluentD is a great idea for...

I am not sure that adding the IP address of the headnode as an annotation is a good practice. Especially that IP address are ephemeral and can change. Plus the...

@davidxia for the GKE case, when you need external access to the cluster, an alternative is to use a `K8s Service` that is of type `load-balancers or NodePort`. Exposing an...

@DmitriGekhtman , I agree, the IP is a runtime generated value, and having it in the status field makes sense to me.

why does the first structure have 2 API folders?

yes today we can add/remove workers from the worker groups. We can remove random workers by changing the replicas or remove specific ones by specifying the pod name in the...

Hi Ricardo, Great to see you hear :) Sure there are many of us that are very comfortable presenting KubeRay. One thing to note here, is while Ray itself is...

@DmitriGekhtman we have prepared slides for KubeRay with @Jeffwan that we presented earlier. We can sync with you (on slack), and co-present the final slides.

I think shielding the user from the init-container logic has its advantages. We can have **the Operator add the init-container** in the pod. (We need to make sure here that...

That's an interesting use case. What constitutes resource constraints? if a cluster needs 10 replicas (worker pods) but only 9 can be scheduled, is this a resource constraint? or if...