Skip to content

chore: update golang.google.org/grpc dep from v1.71.1 to v1.72.0 #777

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 6 commits into from
May 5, 2025

Conversation

shaneutt
Copy link
Member

@shaneutt shaneutt commented May 2, 2025

This updates our golang.google.org/grpc dependency from v1.71.1 to v1.72.0.

This update includes a change that is not backwards compatible for those who were providing their own implementation of the HealthServer interface, as it adds a new List() method to the interface. This is why #722 was failing.

Technically, I don't think Kubernetes calls this List() method anywhere yet, but when I reviewed the proposal for this change I did not see any sign that there was any expectation that it was reasonable to return UNIMPLEMENTED status code for this method like it is for the Watch() method. So as to abide by the specification I went with properly implementing it.

This PR implements the List() method for both the bbr and the epp, but it also adds the distinct ext_proc gRPC service name to the health check (previously, we were just accepting any health check as if it was for the ext_proc service) so that the list method can work as intended and enumerate the named services. I confirmed that Kubernetes handles the SERVICE_UNKNOWN response properly (in fact you can see here that it simply gets mad about any response other than SERVING, and that behavior seems appropriate here).

I made sure to test this locally by building an image, deploying it with an implementation and testing inference requests. Things looked 👍

#722 is no longer needed with this PR, and can be closed.

Release Notes

* bbr and epp gRPC service names are now explicit for health checks (before, you could provide any service name, even one that didn't exist). Health checks in deployments will now need to target the envoy.service.ext_proc.v3.ExternalProcessor service explicitly going forward or health checks will fail.

@k8s-ci-robot k8s-ci-robot requested review from ahg-g and nirrozenbaum May 2, 2025 18:26
@k8s-ci-robot k8s-ci-robot added cncf-cla: yes Indicates the PR's author has signed the CNCF CLA. size/L Denotes a PR that changes 100-499 lines, ignoring generated files. labels May 2, 2025
Copy link

netlify bot commented May 2, 2025

Deploy Preview for gateway-api-inference-extension ready!

Name Link
🔨 Latest commit c4b8a53
🔍 Latest deploy log https://app.netlify.com/sites/gateway-api-inference-extension/deploys/68150e5f943c660008ee3ab7
😎 Deploy Preview https://deploy-preview-777--gateway-api-inference-extension.netlify.app
📱 Preview on mobile
Toggle QR Code...

QR Code

Use your smartphone camera to open QR code link.

To edit notification comments on pull requests, go to your Netlify site configuration.

@@ -7,6 +7,7 @@
bin/*
Dockerfile.cross
artifacts
main
Copy link
Collaborator

Choose a reason for hiding this comment

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

musing: Wondering if this will bite someone later. I dont expect us to have a main dir or file at root. Maybe this is preferred

Copy link
Member Author

@shaneutt shaneutt May 2, 2025

Choose a reason for hiding this comment

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

This is the default binary a go build emits, so it was just buggin me because I was building the binaries locally and didn't want it to try and check that in. Wasn't really thinking anyone would actually add this at any point, but if you strongly prefer we don't add this though, no sweat I'll remove it.

@shaneutt shaneutt requested a review from kfswain May 2, 2025 18:42
@shaneutt
Copy link
Member Author

shaneutt commented May 2, 2025

/label tide/merge-method-rebase

@k8s-ci-robot k8s-ci-robot added the tide/merge-method-rebase Denotes a PR that should be rebased by tide when it merges. label May 2, 2025
@@ -53,12 +53,12 @@ spec:
livenessProbe:
grpc:
port: 9003
service: inference-extension
service: envoy.service.ext_proc.v3.ExternalProcessor
Copy link
Contributor

Choose a reason for hiding this comment

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

what does this represent, would it matter if we remove it?

Copy link
Contributor

Choose a reason for hiding this comment

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

I guess this should match the name in the health check routine, otherwise we return HealthCheckResponse_SERVICE_UNKNOWN.

Wondering why we want to take a dependency on what the service name as defined in https://github.com/envoyproxy/go-control-plane/blob/main/envoy/service/ext_proc/v3/external_processor_grpc.pb.go#L135 when we could control that on our end?

Copy link
Member Author

Choose a reason for hiding this comment

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

To my understanding the gRPC service might not be able to change because ext_proc expects this specific one by name to be present for its requests. At least by default, maybe that's configurable somewhere (on the Envoy side)? In any case, this is the actual service name we've been using since the beginning: if you turn on reflection on the current gRPC server and do a list, this is the exposed service that Envoy is expecting.

@shaneutt shaneutt requested a review from ahg-g May 5, 2025 11:32
@ahg-g
Copy link
Contributor

ahg-g commented May 5, 2025

/lgtm
/approve

@k8s-ci-robot k8s-ci-robot added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label May 5, 2025
@k8s-ci-robot
Copy link
Contributor

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: ahg-g, shaneutt

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

The pull request process is described here

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

@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 5, 2025
@k8s-ci-robot k8s-ci-robot merged commit a305927 into kubernetes-sigs:main May 5, 2025
8 checks passed
@shaneutt shaneutt deleted the grpc-update branch May 5, 2025 15:46
@kfswain kfswain mentioned this pull request May 6, 2025
@smarterclayton
Copy link
Contributor

This was reverted in #788 as it caused failures at runtime (see linked issue for details)

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. lgtm "Looks good to me", indicates that a PR is ready to be merged. size/L Denotes a PR that changes 100-499 lines, ignoring generated files. tide/merge-method-rebase Denotes a PR that should be rebased by tide when it merges.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants