Operating System: Ubuntu 20.04
Kubernetes version: 1.21.5
Kubespray version: 2.17.1
Zabbix Version: 6.0.4

Could you check the host and port availability from Zabbix proxy side? Do you have running kube-apiserver pod?

Alexey, yes, the pod kube-apiserver is running. Host and port are available; also, other units work well. Also, the template "Kubernetes nodes by HTTP" works without problems.

Did you try to increase timeout then?

Yes, of course. With the same result.

Is kube-state-metrics running? Please check endpoint of the service.

Yes, it is running (from zabbix helm chart). The endpoit is zabbix-kube-state-metrics (as it is in the template). Also, I mentioned that I can get to the web-interface of kube-state-metrics through lens (I think it's using port-forwarding, but the most important - yes, kube-state-metrics are functioning).

Actually, I also tried to install kube-state-metrics apart, and it neither worked. It was giving the same result.

in this case please enable debug for Zabbix proxy, check again, then attach debug log to the issue.

As I mentioned, this is a helm installation. If I understand right, you need logs from zabbix-proxy pod. I tried to see how to increase its logs here: https://git.zabbix.com/projects/ZT/repos/kubernetes-helm/browse?at=refs%2Fheads%2Frelease%2F6.0

But there's nothing. Then I just tried to run a command:
zabbix_proxy -R log_level_increase
inside the container. Then I run a couple of times the check, and uploaded the container logs to the text file.

I'm not so sure if you really need the zabbix-proxy deployment logs, as the monitoring is going from the zabbix server to kubernetes API.

Also I decided to attach the logs of zabbix-kube-state-metrics. Maybe it would help.

Proxy: proxy.log

Kube-state-metrics: kube_state_metrics.log

I think the proxy is misconfigured. I do not think it is available for Zabbix server (Administration->Proxies->Last seen):

** Updating '/etc/zabbix/zabbix_proxy.conf' parameter "Server": '127.0.0.1'...updated
** Updating '/etc/zabbix/zabbix_proxy.conf' parameter "ServerPort": '10051'...added

Also the host you added is monitoring by proper "zabbix-proxy" proxy?

> I do not think it is available for Zabbix server

Neither I do. I don't know, I didn't configure it, I just did what documentation says: installed the helm and deleted {{securityContext. }}But anyway, what should it do with a proxy, if it's an HTTP monitoring from the zabbix server by kubernetes API? Also, the insteresting thing is that everything else works without problems.

Also, please note the warnings/errors in kube-state-metrics logs. Can't it be a reason of the problem? As I mentioned, I use kubernetes v 1.25 and 1.26

> the host you added is monitoring by proper "zabbix-proxy" proxy?

This "host" is a kubernetes cluster. As I already mentioned, I've installed it following the documentation from here: https://www.zabbix.com/integrations/kubernetes

Besides other things, it installs a zabbix-proxy deployment. I'm not sure if zabbix-proxy is critical, and if I should configure it at all - the documentation doesn't mention it. I think the problem is not related to zabbix-proxy.

If you mean that you need zabbix logs, I could install a new proxy on a virtual machine, route the Kubernetes Cluster monitoring there and give you the result. Would you like me to do this? Or should I configure the proxy from the helm in another way not mentioned in the documentation?

It is documented:

Change the values according to the environment in the file $HOME/zabbix_values.yaml.

So you need to configure Zabbix server location before installing helm.

did you specify Zabbix server IP / DNS after my suggestion?

Hello again!

Sorry, I didn't have a chance to check it before. Today I could finally get it working.

There are a couple of things I didn't understand from the documentation:

1) It's necessary to monitor the cluster through zabbix-proxy (as the template is called "Kubernetes cluster state by HTTP", I thought it's enough to have a http access to it from the zabbix-server);

2) It's necessary to change values.yaml, at least writing the server address and the proxy name.

We can consider the issue closed; but I'd like to collaborate a little. Can I write a document based on the official instruction, but adding the proxy configuration steps? Or I could write a how-to about the configuration. I can just make a document and attach it here, then you can add it to the official documentation if you like. Do you think it's a good idea, or it's not necessary, to change the documentation? Thanks in advance.

1. Yes, if you do not want to monitor through proxy, you have to open metrics / objects monitoring outside of Kubernetes.
2. Of course

Hm... i do not think it is neccessary, because if you install Zabbix proxy it is always needed to configure it.