shrekris-anyscale

> let's just keep the first two benchmark and select one from deployment graph because That sounds good– I kept [`Deployment Graph: Ensemble`](https://github.com/ray-project/ray/blob/master/release/serve_tests/workloads/deployment_graph_wide_ensemble.py) based on our offline conversation.

Hi @zoltan-fedor, thanks for digging into this issue! I agree that it would be helpful for the docs to discuss the metrics that show up via HTTP requests vs via...

> Sure, I am happy to raise the two tickets and also make a PR for https://github.com/ray-project/ray/pull/1. Thanks that would be great! We really appreciate the help. > My only...

> Anyone new reading this page quick start might takeaway, from the example demonstrated, that this is all about external pre-trained models and how to use them with or deploy...

Production guide link: https://ray--27747.org.readthedocs.build/en/27747/serve/production.html

We had a hypothesis that the user was running into #539, which was fixed by #540. The user upgraded Kuberay to master (which contains #540), but they could still reproduce...

Hi @kevin85421, this is still an issue, but I'm not sure if it's caused by Ray Serve itself or by KubeRay. It's somewhat mitigated by [this Ray change](https://github.com/ray-project/ray/pull/29534), but I...

We've made more progress on this issue. [#33384](https://github.com/ray-project/ray/pull/33384) will further reduce any downtime while the worker nodes are down. That change should ensure minimal downtime when this issue happens. After...

@sihanwang41 is this issue resolved by your change in #1014?

Great! @kevin85421 this issue should be resolved, so I'm closing it.