I am using metricbeat to gather information about my kubernetes cluster. I want to build my own status dashboard to see if my deployments / statefulsets in kubernetes are healthy and I want to be able to dig deeper and show performance values of containers and pods.
Currently my dashboard is looking like this. (created with enhanced-table plugin)
Is it possible to achieve the same in TSVB (except for the filter panel)? Doing grouping on multiple levels and coloring a text cell based on a status (calculation between two fields).
I want to add pod and container usage statistics. When I check the events in elasticsearch I have following to offer:
My table on the bottom is gathering information from metricset.nae: state_pod:
Overall pod metrics can be found in metricset.name: pod:
And detailed container metrics can be found in metricset.name: contianer
Marked in green are metrics / values I would like to show.
I want to be able to filter for example for a logical pod name by clicking on a table entry or in an object of tsvb.
By clicking there I want to see every event which has the pod name value in the fields:
kubernetes.pod.name or at best
At very best I want to also show events where field
kubernetes.*.pod_name is not existing.
Can this somehow be achieved with kibana?
In my opinion it would be much, much easier if same information is always stored in the same field. Store the kubernetes.pod.name always in this field, regardless my event is state_pod, container, pod or whatever.
Same issue I am encountering if i filter for kubernetes.node.name. If I set this filter no pod or container will be shown, because the field (with same meaning) is named differently.