Research Node conditions metrics reported in KSM 2.0 about nri-kubernetes

[roobre]

LGTM, seems like conditions are good to go with KSM 2.0.
Also good news that Node Problem Detector integrates well with KSM!

[paologallinaharbur]

Keep in mind to verify that we can alert on strings (true/false/unknown) without issues otherwise we can use (0/1/2) or different approaches

[invidian]

Looking from the changelog, it seems in KSM 2.0 there was no change to node condition metrics.

kube_node_status_condition metric is considered stable.

Metrics from a single node with Node Problem Detector looks like following:

$ $ curl -s localhost:8080/metrics | grep condition | grep kube_node_status_condition | grep foo
kube_node_status_condition{node="foo",condition="CorruptDockerOverlay2",status="true"} 0
kube_node_status_condition{node="foo",condition="CorruptDockerOverlay2",status="false"} 1
kube_node_status_condition{node="foo",condition="CorruptDockerOverlay2",status="unknown"} 0
kube_node_status_condition{node="foo",condition="KernelDeadlock",status="true"} 0
kube_node_status_condition{node="foo",condition="KernelDeadlock",status="false"} 1
kube_node_status_condition{node="foo",condition="KernelDeadlock",status="unknown"} 0
kube_node_status_condition{node="foo",condition="ReadonlyFilesystem",status="true"} 0
kube_node_status_condition{node="foo",condition="ReadonlyFilesystem",status="false"} 1
kube_node_status_condition{node="foo",condition="ReadonlyFilesystem",status="unknown"} 0
kube_node_status_condition{node="foo",condition="NetworkUnavailable",status="true"} 0
kube_node_status_condition{node="foo",condition="NetworkUnavailable",status="false"} 1
kube_node_status_condition{node="foo",condition="NetworkUnavailable",status="unknown"} 0
kube_node_status_condition{node="foo",condition="MemoryPressure",status="true"} 0
kube_node_status_condition{node="foo",condition="MemoryPressure",status="false"} 1
kube_node_status_condition{node="foo",condition="MemoryPressure",status="unknown"} 0
kube_node_status_condition{node="foo",condition="DiskPressure",status="true"} 0
kube_node_status_condition{node="foo",condition="DiskPressure",status="false"} 1
kube_node_status_condition{node="foo",condition="DiskPressure",status="unknown"} 0
kube_node_status_condition{node="foo",condition="PIDPressure",status="true"} 0
kube_node_status_condition{node="foo",condition="PIDPressure",status="false"} 1
kube_node_status_condition{node="foo",condition="PIDPressure",status="unknown"} 0
kube_node_status_condition{node="foo",condition="Ready",status="true"} 1
kube_node_status_condition{node="foo",condition="Ready",status="false"} 0
kube_node_status_condition{node="foo",condition="Ready",status="unknown"} 0