Skip to content

grpc: do not require unique hostnames with http - #4598

Merged
k8s-ci-robot merged 1 commit into
kubernetes-sigs:mainfrom
howardjohn:spec/grpc-http-unique
May 9, 2026
Merged

k8s-ci-robot merged 1 commit into
kubernetes-sigs:mainfrom
howardjohn:spec/grpc-http-unique

Conversation

@howardjohn

Copy link
Copy Markdown
Contributor

This part of the spec is not implemented by a large number of implementations. That I know of:

  • Istio
  • Kgateway
  • Envoy Gateway
  • Contour
  • Agentgateway

I can only speak for the ones I am involved in, but this is not an accident; this was an intentional diversion from the recommendation here, as this does not meet users needs. HTTP and gRPC are entirely compatible to be on the same hostname, and quite common to do so even if the gRPC maintainers does not consider it a best practice (I am not sure how this is even true, given this opinion comes from Google who hosts HTTP and gRPC on the same googleapis.com?? maybe I am missing something). Instead, this forces users into an arbitrary restriction and then forces them to rewrite all of their routes from GRPCRoute to HTTPRoute which is a regression in usability.

The original GEP:

If at some point in the future, demand for this use case increases and we have
reason to believe that the feasibility of implementation has improved, this
would be a backward compatible change.

Given the reality of the current state of implementations, I believe the demand for this has been met to change the spec.

See also https://kubernetes.slack.com/archives/CR0H13KGA/p1770320071586889

What type of PR is this?

/kind documentation

What this PR does / why we need it:

Which issue(s) this PR fixes:

Fixes #

Does this PR introduce a user-facing change?:

Previously, implementations were required to reject GRPCRoute and HTTPRoutes on the same hostname (however, few implementations did this). Now, implementations may optionally do this, or allow them to coincide.

This part of the spec is not implemented by a large number of
implementations. That I know of:
* Istio
* Kgateway
* Envoy Gateway
* Contour
* Agentgateway

I can only speak for the ones I am involved in, but this is not an
accident; this was an intentional diversion from the recommendation
here, as this does not meet users needs. HTTP and gRPC are entirely
compatible to be on the same hostname, and quite common to do so even if
the gRPC maintainers does not consider it a best practice (I am not sure
how this is even true, given this opinion comes from Google who hosts
HTTP and gRPC on the same `googleapis.com`?? maybe I am missing
something). Instead, this forces users into an arbitrary restriction and
then forces them to rewrite all of their routes from GRPCRoute to
HTTPRoute which is a regression in usability.

The original GEP:

```
If at some point in the future, demand for this use case increases and we have
reason to believe that the feasibility of implementation has improved, this
would be a backward compatible change.
```

Given the reality of the current state of implementations, I believe the
demand for this has been met to change the spec.

See also https://kubernetes.slack.com/archives/CR0H13KGA/p1770320071586889
@k8s-ci-robot k8s-ci-robot added release-note Denotes a PR that will be considered when it comes time to generate release notes. kind/documentation Categorizes issue or PR as related to documentation. labels Feb 24, 2026
@k8s-ci-robot k8s-ci-robot added size/S Denotes a PR that changes 10-29 lines, ignoring generated files. cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. labels Feb 24, 2026
Comment on lines +46 to +47
Implementations that support GRPCRoute may enforce uniqueness of
hostnames between `GRPCRoute`s and `HTTPRoute`s. That is, if a route (A) of type `HTTPRoute` or

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

As the uniqueness is not currently enforced within a Route type (as raised in #4303), this relaxation seems reasonable to me and could be tightened again in the future.

@rikatz

rikatz commented May 8, 2026

Copy link
Copy Markdown
Member

yeah, I agree with this as well. It is desired to not mix the hostnames on HTTPRoute and GRPCRoute, but it is not like this needs an enforcement.

/approve

@snorwin leaving the lgtm for you.

Thanks @howardjohn

@k8s-ci-robot k8s-ci-robot added the approved Indicates a PR has been approved by an approver from all required OWNERS files. label May 8, 2026

@snorwin snorwin left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

/lgtm

@howardjohn thanks.

@k8s-ci-robot k8s-ci-robot added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label May 9, 2026
@k8s-ci-robot

Copy link
Copy Markdown
Contributor

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: howardjohn, rikatz, snorwin

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Details Needs approval from an approver in each of these files:

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

approved Indicates a PR has been approved by an approver from all required OWNERS files. cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. kind/documentation Categorizes issue or PR as related to documentation. lgtm "Looks good to me", indicates that a PR is ready to be merged. release-note Denotes a PR that will be considered when it comes time to generate release notes. size/S Denotes a PR that changes 10-29 lines, ignoring generated files.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants