How can I get node's memory with kubernetes-client lib in python

Question

Could not find how can I obtain information about memory consumption on kubernetes node with kubernetes-client library in Python. I know how to obtain this information with kubectl command.

Could you provide me with a piece of code how can I do this via the lib?

Thank you.

Asked By: Stanislav Hordiyenko

||

Source

Answer 1

Getting started

kubernetes-client contains pieces of code that you need:

from kubernetes import client, config


def main():
    # Configs can be set in Configuration class directly or using helper
    # utility. If no argument provided, the config will be loaded from
    # default location.
    config.load_kube_config()

    v1 = client.CoreV1Api()
    print("Listing pods with their IPs:")
    ret = v1.list_pod_for_all_namespaces(watch=False)
    for i in ret.items:

        print("%st%st%s" %
              (i.status.pod_ip, i.metadata.namespace, i.metadata.name))


if __name__ == '__main__':
    main()

Method

For get pod metrics, try this beta class:
V2beta2PodsMetricStatus

Metrics

According to k8s documentation Your metrics supposed to be:

kube_pod_container_resource_limits_memory_bytes
kube_pod_container_resource_requests_memory_bytes

Answered By: Yasen

Answer 2

After a long investigation and some good advice (use --v=10 flag in kubectl command) from the kubernetes-client repository on GitHub I managed to find two resources where I can fetch CPU/memory consumption metrics that I was interested in:

https://your.kube-url.com/api/v1/namespaces/kube-system/services/http:heapster:/proxy/apis/metrics/v1alpha1/namespaces/default/pods
https://your.kube-url.com/api/v1/namespaces/kube-system/services/http:heapster:/proxy/apis/metrics/v1alpha1/nodes

I couldn’t find any code examples how to do the same with kubernetes-client library that is why will build my own wrappers.

Answered By: Stanislav Hordiyenko

Answer 3

Belated answer, but since this frustrated me so much and I wished there had been a proper answer to this question so I didn’t have to go digging for it I’m writing one now to make it easier on the next poor sap.

Here is how you can get memory for nodes:

from kubernetes import client

custom_api = client.CustomObjectsApi(client.ApiClient()) 
response = custom_api.list_cluster_custom_object("metrics.k8s.io", "v1beta1", "nodes")

 for node in response['items'];
     print(f"node {node['metadata']['name']} has available memory of {node['usage']['memory']}")

If you get an unauthorized response when you try to run this your need to update your clusterRole and clusterRoleBinding to authorize this. Your need the ‘list’ action on the resource ‘nodes’ for the v1beta1 api

Answered By: dsollen

How can I get node's memory with kubernetes-client lib in python

Question:

Answers:

Getting started

Method

Metrics