Now we have one view for the CPU usage and for the Memory usage:
In Kibana code is checked availability of 2 metrics: kubernetes.pod.cpu.usage.node.pct and kubernetes.pod.cpu.usage.limit.pct
kubernetes.pod.cpu.usage.limit.pct represents the pod cpu usage as a ratio of the pod cpu limits
kubernetes.pod.cpu.usage.node.pct represents the pod cpu usage as a ratio of the node cpu allocatable (relyin on the metric kubernetes.node.cpu.allocatable.cores)
On the same waffle board are mixed 2 metrics: usage.limit.pct and usage.node.pct, this behavior is misleading.
Describe the feature:
My proposal is to split the views (CPU usage and Memory usage) and have 2 view per resource to avoid this confusion and avoid mixing up different metrics on the same board.
For cpu:
CPU usage per Pod limit - that relies on kubernetes.pod.cpu.usage.limit.pctCPU usage per Node limit - kubernetes.pod.cpu.usage.node.pct
memory:
Memory usage per Pod limit - kubernetes.memory.cpu.usage.limit.pctMemory usage per Node limit - kubernetes.memory.cpu.usage.node.pct
Now we have one view for the
CPU usage
and for theMemory usage
:In Kibana code is checked availability of 2 metrics:
kubernetes.pod.cpu.usage.node.pct
andkubernetes.pod.cpu.usage.limit.pct
kubernetes.pod.cpu.usage.limit.pct
represents the pod cpu usage as a ratio of the pod cpu limitskubernetes.pod.cpu.usage.node.pct
represents the pod cpu usage as a ratio of the node cpu allocatable (relyin on the metrickubernetes.node.cpu.allocatable.cores
)The same applies to the memory metrics
On the same waffle board are mixed 2 metrics:
usage.limit.pct
andusage.node.pct
, this behavior is misleading.Describe the feature: My proposal is to split the views (
CPU usage
andMemory usage
) and have 2 view per resource to avoid this confusion and avoid mixing up different metrics on the same board. For cpu:CPU usage per Pod limit
- that relies onkubernetes.pod.cpu.usage.limit.pct
CPU usage per Node limit
-kubernetes.pod.cpu.usage.node.pct
memory:Memory usage per Pod limit
-kubernetes.memory.cpu.usage.limit.pct
Memory usage per Node limit
-kubernetes.memory.cpu.usage.node.pct