Fix 15877 and bump up gRPC from v1.52.0 to v1.57.0

[chaochn47]

Fix #16290

Pending to evaluate potential breaking changes mentioned in #15145 (comment).

Please read https://github.com/etcd-io/etcd/blob/main/CONTRIBUTING.md#contribution-flow.

[chaochn47]

One flaky robustness test execution with LazyFS link. Retry should help.

traffic.go:106: Requiring minimal 100.000000 qps for test results to be reliable, got 99.438205 qps

[ahrtr]

This PR seems to be almost identical to #16301

[liggitt]

This PR seems to be almost identical to #16301

this picks up the genproto version I cared about, so if it's passing presubmits (#16301 is hitting a weird timeout), I'm happy to see this merge instead

[chaochn47]

Pending to evaluate potential breaking changes mentioned in #15145 (comment).

I have reviewed merged changes especially grpc/grpc-go#6471 and grpc/grpc-go#6481. Those should be backward compatible changes until grpc-go completely remove the deprecated API like resolver.State.Addresses. It will be obvious that etcd cannot compile since the symbols will be removed and captured by the CI.

this picks up the genproto version I cared about, so if it's passing presubmits (#16301 is hitting a weird timeout), I'm happy to see this merge instead

Thanks, I will mark the PR as ready to review with the above two arguments.

[wenjiaswe]

lgtm

[ahrtr]

I have reviewed merged changes especially grpc/grpc-go#6471 and grpc/grpc-go#6481. Those should be backward compatible changes until grpc-go completely remove the deprecated API like resolver.State.Addresses.

Confirmed. We still need to remove the dependency on the experimental gRPC API, as tracked in #15145. Let's continue the discussion in that issue separately.

[ahrtr]

LGTM

thx @chaochn47