Zhewei Hu

> Hi @phlax , Is there a plan to make a 1.31 release with this change. We are pulling the 1.31.2 and it is failing. I think release branch 1.31...

I got exact the same issue and the gateway programmed status will be `Unknown` indefinitely. From the istiod log. I think multiple istiods were step over each other > error...

cc @alyssawilk and @mattklein123, any thoughts?

> experiment with preconnect Thank you for the pointer! I went through the [doc](https://www.envoyproxy.io/docs/envoy/latest/api-v3/config/cluster/v3/cluster.proto#config-cluster-v3-cluster-preconnectpolicy:~:text=For%20HTTP/2%2C%20only%20one%20connection%20would%20be%20established%20by%20default%20as%20one%20connection%20can%20serve%20both%20the%20original%20and%20presumed%20follow%2Dup%20stream.), looks like per_upstream_preconnect_ratio will only work for non-multiplexed connections?

> So the problem here is you have a high traffic service, you have a set connection drain time, and all your connections drain around the same time resulting in...

Thank you for the confirmation! No problem at all! I really should highlight the downstream in the description.

This issue is not stale, I will try to work on the implementation.

> These are not the same tests. The HttpConnectionManagerImplTest test validates integration of UHV with HCM, while the Http2HeaderValidatorTest validates logic in the default HTTP/2 header validator. > > /wait-any...

> > > These are not the same tests. The HttpConnectionManagerImplTest test validates integration of UHV with HCM, while the Http2HeaderValidatorTest validates logic in the default HTTP/2 header validator. >...

> > * HeaderValidatorFailTrailersTransformationBeforeResponse > > These do look like the same test. Thank you for the confirmation! And I removed the 2nd test case in this PR.