Advanced Kubectl Commands

  • How to use JSON PATH with kubectl in Kubernetes.

  • Good to practice with JSON PATH on a Kubernetes dataset.

  • Why do we use JSON PATH

    • Useful for large datasets in Kubernetes and for big Production environments - need to see information about 100s of nodes and 1000s of PODs, Deployments and ReplicaSets

      • Will have requirements of printing summary of states of different resources, want to view different fields about resources.

        • kubectl supports JSON PATH to make filtering data across large datasets much easier.

  • How does the kubectl utility work? It is the Kubernetes CLI and allows us to interact with Kubernetes Objects.

  • When running the following:

kubectl get nodes

  • That is directly connecting to the kube-apiserver.

    • The kube-apiserver speaks the JSON language and therefore returns things in a JSON format.

    • The output of the kubectl get nodes command is actually shortened from the JSON PATH output.

      • To print additoinal details, can use kubectl get nodes -o wide

  • A good use of JSON PATH and kubectl - to show nodes and their CPU counts in a tabular format:

    • Also print a list of pods and the images they use.

      • None of the built-in commands can output exactly these.

  • JSON PATH queries allow you to filter the output of a command as you like.

  • How to get started in JSON PATH in kubectl?

    • Identify the kubectl command (kubectl get nodes, kubectl get pods etc)

    • Familiarise with the JSON format - printing the output in JSON - kubectl get nodes -o json / kubectl get pods -o json.

    • Look through the structure of the JSON document and form the JSON PATH query.

      • An example is using .items[0].spec.containers[0].image

    • Then use the query you developed with the same kubectl command. Using the above example:

        kubectl get pods -o=jsonpath='{.items[0].spec.containers[0].image}'

  • Good to use a JSON PATH evaluator like https://jsonpath.com/

  • Play around with it until you have the best JSON PATH query and then add that to the kubectl command.

  • A good example JSON PATH command to provide the names of the nodes:

kubectl get nodes -o=jsonpath='{.items[*].metadata.name}'

* The output would be `master node01`
  • A good example of returning the hardware architecture of the nodes:
kubectl get nodes -o=jsonpath='{.items[*]}.status.nodeInfo.architecture}'
  • The output would be:
amd64 amd64
  • A good command on returning the CPU count of a node:
kubectl get nodes -o=jsonpath='{.items[*].status.capacity.cpu}'
  • The output here would be:
4 4
  • Can merge the queries together to get multiple output, like so:
kubectl get nodes -o=jsonpath='{.items[*].status.capacity.cpu}{.items[*]}.status.nodeInfo.architecture}'
  • The output would be:
master node01 4 4
  • You can make the above output much more pretty though:
{"\n"} - new line
{"\t"} - tab
  • An example is below:
kubectl get nodes -o=jsonpath='{.items[*].status.capacity.cpu}{"\n"}{.items[*]}.status.nodeInfo.architecture}' 

master node01 
4       4

  • You can also use loops with range.

    • We iterate through items in the list and create properties of each item.

  • To get the format we want below:

master 4
node01 4
  • We need to do this pseudocode:
FOR EACH NODE

  PRINT NODE NAME \t PRINT CPU COUNT \n
  
END FOR
  • To specify the FOR EACH statement, we need the range keyword:
'{range.items[*]}'

* The `*` means for each item.
  • To print the node name and cpu, use the same line from earlier:
kubectl get nodes -o=jsonpath='{range.items[*]{.metadata.name} {"\t"} {.status.capacity.cpu} {"\n"}{end}'
  • Can also use JSON PATH when printing custom columns:
kubectl get nodes -o=custom-columns=<COLUMN NAME>:<JSON_PATH>
  • An example of using columns is below:
kubectl get nodes -o=custom-columns=NODE:.metadata.name ,CPU:.status.capacity.cpu
  • This looks liks this:
NODE CPU
master 4
node01 4

  • You can also sort the output based on a property from each item.

  • Examples are below:

kubectl get nodes --sort-by=.metadata.name
kubectl get node s--sort-by=.status.capacity.cpu

Updated: