Unmentioned breaking change in 1.11 for GRPC

[flphvlck]

There is unmentioned breaking change in 1.11 related to GRPC in combination with nginx.ingress.kubernetes.io/configuration-snippet annotation.

This feature came from issue #11250 and pull request #11258.

It would be nice to mention it in release notes.

What happened:
Nginx exits with error:

Error: exit status 1
2024/08/19 12:10:49 [emerg] 60#60: "grpc_read_timeout" directive is duplicate in /tmp/nginx/nginx-cfg658058193:2367
nginx: [emerg] "grpc_read_timeout" directive is duplicate in /tmp/nginx/nginx-cfg658058193:2367
nginx: configuration file /tmp/nginx/nginx-cfg658058193 test failed

This happens because we are using

      nginx.ingress.kubernetes.io/configuration-snippet: |
        grpc_next_upstream_tries X;
        grpc_read_timeout Xs;
        grpc_send_timeout Xs;

In the /tmp/nginx/nginx-cfg658058193 there really are duplicated directives for grpc (default values from annotations and our custom values from configuration-snippet

			# Grpc settings
			grpc_connect_timeout                    5s;
			grpc_send_timeout                       60s;
			grpc_read_timeout                       60s;

			grpc_next_upstream error timeout http_502 http_503 http_504 non_idempotent;
			grpc_next_upstream_tries X;
			grpc_read_timeout Xs;
			grpc_send_timeout Xs;

[flphvlck]

If it's just about editing files in https://github.com/kubernetes/ingress-nginx/tree/main/changelog and the rest is magic behind, I can do it.
something like:

Breaking change
Stop using grpc_connect_timeout, grpc_send_timeout, grpc_read_timeout in nginx.ingress.kubernetes.io/configuration-snippet annotation (otherwise controller crash). These GRPC options (and eventually their default values) are inherited from respective proxy timeouts. Find out more in docs: https://kubernetes.github.io/ingress-nginx/user-guide/nginx-configuration/annotations/#custom-timeouts

[longwuyuan]

/assign @Gacko
/triage accepted

[github-actions]

This is stale, but we won't close it automatically, just bare in mind the maintainers may be busy with other tasks and will reach your issue ASAP. If you have any question or request to prioritize this, please reach #ingress-nginx-dev on Kubernetes Slack.

[soupdiver]

It's month now and just after breaking our deployments I ended up here.
Is there any plan to do something about this?
An updated changelog or so would also help.

[cflmarques]

We also had this exact issue. Any updates?

[strongjz]

We apologize this caused issues, we are working on updating the templating engine in hopes that it would catch issues like this. There hundreds of possible config combinations and we request end to end test for all new features. The major issue with config and server snippets is that we can never tell exactly what a user will have in them, making it near impossible to test that scenario.

We generally do not update release notes, as they are automatically generated during releases.

I suggest moved to using the new feature and not using the config snippets.

/close

[k8s-ci-robot]

@strongjz: Closing this issue.

In response to this:

We apologize this caused issues, we are working on updating the templating engine in hopes that it would catch issues like this. There hundreds of possible config combinations and we request end to end test for all new features. The major issue with config and server snippets is that we can never tell exactly what a user will have in them, making it near impossible to test that scenario.

We generally do not update release notes, as they are automatically generated during releases.

I suggest moved to using the new feature and not using the config snippets.

/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.