Can a kubernetes member verify that this patch is reasonable to test? If so, please reply with "@k8s-bot ok to test" on its own line.

Regular contributors should join the org to skip this step.

While we transition away from the Jenkins GitHub PR Builder plugin, "ok to test" commenters will need to be on the admin list defined in this file.

ok to test

I don't think it was remove by accident. Take a look to #26652 (comment)

@aledbf could you please provide more context?
@crassirostris do you understand the change?

could you please provide more context?

The kubernetes_metadata plugin was commented because the pod starts without the service account and token mounted which avoid the normal initialization of the fluentd process.

#26652 (comment)

Yes, sure, the fact that this plugin is needed is out of question.

But just turning it on doesn't seem like a good idea, since it was turned off not without a reason.

Maybe we should provide this filter plugin with explicit path to cert?

My point is, let's solve this problem

@aledbf Do you have any idea right now as for how to fix the problem, mentioned in #26652 ?

Maybe we should provide this filter plugin with explicit path to cert?

that fixes the issue

Using Daemonset also solves the issue #32088

Do you have any idea right now as for how to fix the problem, mentioned in #26652 ?

No but If you start the fluentd pod manually (adding a new manifest file) after the node is running the service account is mounted in the pod so I think this issue is just a timing problem.

Well, if this pr passes gce suite for elasticsearch logging, sounds reasonable

@aledbf yep, actually I created my own daemonset for fluentd, and it works fine. didn't know the issue related to pure pod.

In tests the old image is used. Let's wait for #32088 to be merged. I'll prioritize it on my list.

@piosz Do you mind If we will add a comment in the td-agent.conf to "enable" the metadata-plugin in a case of daemon-set usage?.

@keglevich3 Yes, absolutely, go ahead, we will be grateful!

Sorry for the delay. I'm pretty close to merge #32088 (hopefully will be done within next few days) and then we can disable move forward with this one.

@piosz done.
JFYI: I already implemented full EFK stack in my private repo: https://github.com/kayrus/elk-kubernetes

@kayrus LGTM, thanks for your efforts!

Just a small thing: could you please bump an image version? I will push it and then you can create another PR using the new image.

@crassirostris done

Jenkins unit/integration failed for commit 8435d19. Full PR test history.

The magic incantation to run this job again is @k8s-bot unit test this. Please help us cut down flakes by linking to an open flake issue when you hit one in your PR.

Jenkins GKE smoke e2e failed for commit 8435d19. Full PR test history.

The magic incantation to run this job again is @k8s-bot cvm gke e2e test this. Please help us cut down flakes by linking to an open flake issue when you hit one in your PR.

Jenkins GCI GKE smoke e2e failed for commit 8435d19. Full PR test history.

The magic incantation to run this job again is @k8s-bot gci gke e2e test this. Please help us cut down flakes by linking to an open flake issue when you hit one in your PR.

@k8s-bot unit test this

Automatic merge from submit-queue

Commit found in the "release-1.5" branch appears to be this PR. Removing the "cherrypick-candidate" label. If this is an error find help to get your PR picked.