Right now I have configured variables inside my Grafana Dashboard for the Groups and the hosts - see image.
How do I manage to monitor individual pods and namespaces for Kubernetes? How do I configure the variables for that?
Documentation on Zabbix-Grafana-Kubernetes is sheer and there are no out-of-the-box dashboards for that.
I Use Zabbix 7.0 and Grafana 4.4.3
First things first, zabbix 7.0 is alpha. You shouldn’t use it and shouldn’t ask for support on it.
Grafana should be only a graphical representation of your already organized Zabbix when you combine the two tools. Try to model your Zabbix with some Tags to filter each namespace and each pod to represent it in another layers of variables at Grafana’s side.
About the docs, honestly… Zabbix isn’t really ready for cloud deployments yet. Maybe 7.0 when fully released would address this issue, so far it’s a great tool for environments like Networking hosts, Bare Metal hosting and a little bit of application monitoring. But that’s not a subject/discussion to focus on this forum. Better try at https://www.zabbix.com/forum
First things first, why is there a Zabbix plugin available in Grafana Cloud then?
I want to use Zabbix because I don’t a million dollars to spend on Prometheus metrics…
Just trying to get some help
All plugins already available at Grafana Plugins - extend and customize your Grafana are ready to be used in Grafana Cloud, just requiring attention to free and paid subscriptions and their limitations.
If you use Gratana and Prometheus onpremise and self manager, you will only have to pay for learning and management time.
You’re focusing on the problems when you should be focusing on the solution, so let’s try again:
Try to model your Zabbix templates and data with some Tags to filter each namespace and each pod. When you got this classification on Zabbix’s side, then you can represent it in another layers of variables using the tags filter at Grafana’s side.
Can you try it like above and give some feedback for us?
My context before was to make you understand why isn’t practical to make it working only on Grafana’s side.
I have configured these tags in Zabbix on all my nodes:
and set these variables (idk if this is the right way):
This doesn’t work tho
No… It’s not configured the right way in Zabbix and it’s not the right way on Grafana… In both tools, they are generic monitoring systems, they don’t understand “namespaces” or “pods” automatically… You’ll need to show in both tools how to work with the data…
You’re currently setting a hardcoded value for the variable. It isn’t feasible or make sense…
- Namespace tag value should be something like the discovered LLD Macro of your namespaces
- pod tag value should be something like the discovered LLD Macro of your pods inside a namespace.
I advise you to make your way in Zabbix, learning about LLD Discoveries and do it till you cleary see the organization of these tags working inside Zabbix first.
Only after it makes sense in Zabbix side, you should move and try to configure Grafana visualization.
BTW, some zabbix docs resources for ya: