Kubernetes Cheat Sheet for day to day DevOps operations?

calendar Jan 15, 2022 · 11 min read  ·
Share on: linkedin copy

Note - This Kubernetes Cheat sheet is really long and detailed. I would highly recommend to use CTRL+F from the browser to search for the specific command.

Table Of Content

  1. Pods
  2. Deployments
  3. Port Forwarding
  4. Service
  5. Nodes
  6. Namespace
  7. Service Accounts
  8. Documentation
  9. Describing Resources
  10. Editing Resources
  11. Deleting Resources
  12. All get commands
  13. Abbreviations / Short forms of resource types
  14. Verbose Kubectl
  15. Cluster
  16. Kubectl Context
  17. Alias
  18. Kubectl imperative(create) vs declarative(apply)


1. Pods

List all pods in namespace

kubectl get pods

1kubectl get pod

or

1kubectl get pod

or

1kubectl get po

View a pod in watch mode

1kubectl get pod <pod> --watch

View all pods in watch mode

1kubectl get pods -A --watch

List sroted pods

1kubectl get pods --sort-by='.status.containerStatuses[0].restartCount'

List pods using a different output

1kubectl get pods -o <json|yaml|wide|custom-columns=...|custom-columns-
2file=...|go-template=...|go-template-file=...|jsonpath=...|jsonpath-file=...>

Examples:

  1. JSON output
1kubectl get pods -o json

or

1kubectl get pods -ojson
  1. Wide output:
1kubectl get pods -o wide
  1. Custom columns:
1kubectl get pods -o custom-columns='DATA:spec.containers[*].image'

or

1kubectl get pods -o custom-columns='DATA:spec.containers[*].volumeMounts'

or

1kubectl get pods -o custom-columns='DATA:metadata.*'


Formatting output

To output details to your terminal window in a specific format, add the -o (or --output ) flag to a supported kubectl command (source: Kubernetes docs)

Output format Description
-o=custom-columns = spec Print a table using a comma separated list of custom columns
-o=custom-columns-file= filename Print a table using the custom columns template in the filename file
-o=json Output a JSON formatted API object
-o=jsonpath=template Print the fields defined in a jsonpath expression
-o=jsonpath-file= filename Print the fields defined by the jsonpath expression in the file
-o=name Print only the resource name and nothing else
-o=wide Output in the plain-text format with any additional information, and for pods, the node name is included
-o=yaml Output a YAML formatted API object

List all pods in a namespace

1kubectl get pods -n <namespace>

or

1kubectl -n <namespace> get pods

or

1kubectl --namespace <namespace> get pods

List all pods in all namespaces

1kubectl get pods --all-namespaces

or

1kubectl get pods -A


Create from an image

1kubectl run <pod> --generator=run-pod/v1 --image=<image>

In the following cheatsheet, we will be using images such as nginx or busybox.

Example:


1kubectl run nginx --generator=run-pod/v1 --image=nginx

1kubectl run busybox --generator=run-pod/v1 --image=busybox

Run pod in an interactive shell mode

1kubectl run -i --tty nginx --image=nginx -- sh

Run a command after creating a pod

1kubectl run busybox --image=busybox -- sleep 100000

Executing a command in a running pod

1kubectl exec <pod> -- <command>

Or pass stdin to the container in TTY mode:

1kubectl exec -it <pod> -- <command>

Example:

1kubectl exec -it nginx -- ls -lrth /app/


Create a pod: dry run mode (without really creating it)

1kubectl run <pod> --generator=run-pod/v1 --image=nginx --dry-run

Patch a pod

1kubectl patch pod <pod> -p '<patch>'

Example:

1kubectl patch pod <pod> -p '{"spec":{"containers":[{"name":"kubernetes-serve-
2hostname","image":"new image"}]}}'

Another example:

1kubectl patch pod valid-pod --type='json' -p='[{"op": "replace", "path":
2"/spec/containers/0/image", "value":"new image"}]'

Create from a YAML file

1kubectl create -f pod.yaml

Export YAML from the dry run mode

1kubectl run nginx --generator=run-pod/v1 --image=nginx --dry-run -o yaml

Create from STDIN

 1cat <<EOF | kubectl create -f -
 2apiVersion: v1
 3kind: Pod
 4metadata:
 5  name: nginx-pod
 6spec:
 7    containers:
 8    - name: nginx
 9      image: nginx:latest
10EOF


Create multiple resources from STDIN

 1cat <<EOF | kubectl create -f -
 2apiVersion: v1
 3kind: Pod
 4metadata:
 5  name: nginx-pod
 6spec:
 7  containers:
 8  - name: nginx
 9    image: nginx:latest
10---
11apiVersion: v1
12kind: Pod
13metadata:
14  name: busybox
15spec:
16  containers:
17    - name: busybox
18      image: busybox
19      args:
20      - sleep
21      - "100"

Create in a namespace

1kubectl run nginx --generator=run-pod/v1 --image=nginx -n <namespace>

Create in a namespace from a file

1kubectl create -f pod.yaml -n <namespace>

Delete pods

1kubectl delete pod/<pod>

or

1kubectl delete pod <pod>

If you create the pod from a file, you can also use:

1kubectl delete -f pod.yaml

To force deletion:

1kubectl delete pod <pod> --grace-period=0 --force

Get pod logs

1kubectl logs <pod>

or

Sometimes a pod contains more than 1 container. You need to filter the output to get logs for a specific container(s)

1kubectl logs <pod> -c <container>

To follow the logs output (tail -f):

1kubectl logs -f <pod>

If you need to output the logs for all pods with a label

1kubectl logs -l <label_name>=<label_value>

Example:

1kubectl logs -l env=prod

You can also view logs in a multi container case with labels:

1kubectl logs -l <label_name>=<label_value> -c <container>

Or view all cotainers logs with a given label:

1kubectl logs -f -l <label_name>=<label_value> --all-containers

List all container id of init container of all pods

1kubectl get pods --all-namespaces -o jsonpath='{range
2.items[*].status.initContainerStatuses[*]}{.containerID}{"\n"}{end}' | cut -d/ -
3f3

Show metrics for a given pod

1kubectl top pod <pod>

Show metrics for a given pod and all its containers

1kubectl top pod <pod> --containers

2. Deployments

Create a deployment

1kubectl run <deployment> --image=<image>

or

1kubectl create deployment <deployment> --image=<image>

Create a deployment with a predefined replica number

1kubectl run <deployment> --image=<image> --replicas=<number>

Create a deployment with a predefined replica number and opening a port

1kubectl run <deployment> --image=<image> --replicas=<replicas> --port=<port>

Example:

1kubectl run nginx --image=nginx --replicas=2 --port=80

Note: The default generator for kubectl run is --generator=deployment/apps.v1 .

Note: --generator=deployment/apps.v1 is deprecated and will be removed in future versions. Use kubectl run --generator=run-pod/v1 or kubectl create instead.


Create a deployment with a predefined replica number,opening a port and exposing it

1kubectl run nginx --image=nginx --replicas=2 --port=80 --expose

Get a deployment

1kubectl get deploy <deployment>

Watch a deployment

1kubectl get deployment <deployment> --watch

or

1kubectl get deployment <deployment> -w

Or using a shorter version:

1kubectl get deploy <deployment> -w

Or even the longer one:

1kubectl get deployments.apps <deployment> --watch

List all deployments

Same as listing pods, you have multiple options from namespace to output formatters:

1kubectl get deploy -n <namespace>
2
3kubectl get deploy --all-namespaces
4kubectl get deploy -A
5
6kubectl get deploy -oyaml
7kubectl get deploy -owide

Update the image

Rolling update "nginx" containers of "nginx" deployment, updating the image:

1kubectl set image deployment/nginx nginx=nginx:1.9.1

Rolling update "api" containers of "backend" deployment, updating the image:

1kubectl set image deployment/backend api=image:v2

Scale a deployment

1kubectl scale --replicas=5 deployment/<deployment>

Note: You can use a shorter version:

1kubectl scale --replicas=5 deploy/<deployment>

Dry run and YAML output

1kubectl run nginx --image=nginx --replicas=2 --port=80 --dry-run -o yaml

Create a deployment from a file

1kubectl apply -f deployment.yaml

Edit a deployment

1kubectl edit deployment/<deployment>

Rollback deployment

After editing your deployment, you had an error, a solution can be rolling back to the old deployment status:

1kubectl rollout undo deployment <deployment>

Get rollout history

You can check the rollout history:

1kubectl rollout history deployment <deployment>

1kubectl rollout history deployment <deployment>

Example:

1kubectl rollout history deployment nginx

gives you:

1REVISION CHANGE-CAUSE
22         kubectl set image deployment/nginx nginx=nginx:1.9.1 --record=true
33         <none>

Roll back to a previous revision

Using the information from the rollout history, we can get back our deployment to a given revision:

1kubectl rollout undo deployment <deployment> --to-revision=<revision>

Example:

1kubectl rollout undo deployment nginx --to-revision=2

Execute deployment rollout operations

1kubectl rollout status deployment <deployment>
2kubectl rollout pause deployment <deployment>
3kubectl rollout resume deployment <deployment>

3. Port Forwarding

Choosing localhost port

1kubectl port-forward deployment <deployment> <locahost-port>:<deployment-port>
2kubectl port-forward pod <pod> <locahost-port>:<pod-port>

Example: Forward to localhost 8090 from pod 6379:

1kubectl port-forward redis 8090:6379

Listening on the same port

1kubectl port-forward pod <pod> <port>

Example: Listen on ports 8000 and 9000 on localhost, forwarded from the same ports in the pod (8000 and 9000)

1kubectl port-forward pod nginx 8000 9000

Listen on a random port locally

1kubectl port-forward pod <pod> :<pod-port>

Example:

1kubectl port-forward pod nginx :80

Listen on port on localhost + another IP

1kubectl port-forward --address localhost,<IP.IP.IP.IP> pod <pod> <locahost-
2port>:<pod-port>

Example:

1kubectl port-forward --address localhost,10.10.10.1 pod redis 8090:6379

Listen on a forwarded port on all addresses

1kubectl port-forward --address 0.0.0.0 pod <pod> <hosts-port>:<pod-port>

4.Services

Create a service

1kubectl create service <clusterip|externalname|loadbalancer|nodeport> <service>
2[flags] [options]>

Examples:

1kubectl create service clusterip myclusterip --tcp=5678:8080
2kubectl create service loadbalancer myloadbalancer --tcp=80

You can use svc instead of service


Delete service(s)

1kubectl delete service myclusterip
2kubectl delete service myloadbalancer
3
4kubectl delete svc myclusterip
5kubectl delete svc myloadbalancer

or

1kubectl delete service myclusterip myloadbalancer

Describe a service

1kubectl describe service <service>

5.Nodes

Get node

1kubectl get nodes

Get a specific node

1kubectl get nodes <node>

Show node metrics

1kubectl top node <node>

Get external IPs of cluster nodes

1kubectl get nodes -o jsonpath='{.items[*].status.addresses[?
2(@.type=="ExternalIP")].address}'

Describe commands with verbose output

1kubectl describe nodes <node>

Check which nodes are ready

1JSONPATH='{range .items[*]}{@.metadata.name}:{range @.status.conditions[*]}
2{@.type}={@.status};{end}{end}' && kubectl get nodes -o jsonpath="$JSONPATH" |
3grep "Ready=True"

Mark a node as unschedulable

1kubectl cordon <node>

Drain a node for maintenance

1kubectl drain <node>

Mark a node as schedulable

1kubectl uncordon <node>

6.Namespaces

List namespaces

1kubectl get namespaces

or

1kubectl get ns

List or describe a namespace

1kubectl get namespace <namespace>
2kubectl describe namespace <namespace>

Create namespace

1kubectl create namespace <namespace>

or

1kubectl create -f namespace.yaml

or

1cat <<EOF | kubectl create -f -
2apiVersion: v1
3kind: Namespace
4metadata:
5  name: mynamespace
6EOF

Delete namespace

1kubectl delete namespace <namespace>

or

1kubectl delete -f namespace.yaml

7. Service accounts

List service accounts

1kubectl get serviceaccounts

or

1kubectl get sa

Get a service account

1kubectl get serviceaccount <serviceaccount>

or

1kubectl get serviceaccounts <serviceaccount>

or

1kubectl get sa <serviceaccount>

or

1kubectl get sa/<serviceaccount>

Create a service account

1kubectl create serviceaccount <serviceaccount>

Delete a service account

1kubectl delete serviceaccount <serviceaccount>

or

1kubectl delete -f myserviceaccount.yaml

Describe a service account

1kubectl describe serviceaccount <serviceaccount>

Events

List events

1kubectl get events -A

List sorted events

1kubectl get events --sort-by=<JSONPath>

Example: Sorted by timestamp

1kubectl get events --sort-by=.metadata.creationTimestamp

List formatted events

1kubectl get events -o <json|yaml|wide|custom-columns=...|custom-columns-
2file=...|go-template=...|go-template-file=...|jsonpath=...|jsonpath-file=...>

Example:

1kubectl get events -owide

8.Documentation

Get the documentation for pod manifests

1kubectl explain pod

Get the documentation for service manifests

1kubectl explain service

9.Describing resources

1 kubectl describe <resource> <reosurce_name>

Example:

1kubectl describe pod busybox

or

1kubectl describe nodes minikube

Other possible resources you can use with describe :

 1apiservices.apiregistration.k8s.io
 2certificatesigningrequests.certificates.k8s.io
 3clusterrolebindings.rbac.authorization.k8s.io
 4clusterroles.rbac.authorization.k8s.io
 5componentstatuses
 6configmaps
 7controllerrevisions.apps
 8cronjobs.batch
 9csidrivers.storage.k8s.io
10csinodes.storage.k8s.io
11customresourcedefinitions.apiextensions.k8s.io
12daemonsets.apps
13daemonsets.extensions
14deployments.apps
15deployments.extensions
16endpoints
17events
18events.events.k8s.io
19horizontalpodautoscalers.autoscaling
20ingresses.extensions
21ingresses.networking.k8s.io
22jobs.batch
23leases.coordination.k8s.io
24limitranges
25mutatingwebhookconfigurations.admissionregistration.k8s.io
26namespaces
27networkpolicies.extensions
28networkpolicies.networking.k8s.io
29nodes
30persistentvolumeclaims
31persistentvolumes
32poddisruptionbudgets.policy
33pods
34podsecuritypolicies.extensions
35podsecuritypolicies.policy
36podtemplates
37priorityclasses.scheduling.k8s.io
38replicasets.apps
39replicasets.extensions
40replicationcontrollers
41resourcequotas
42rolebindings.rbac.authorization.k8s.io
43roles.rbac.authorization.k8s.io
44runtimeclasses.node.k8s.io
45secrets
46serviceaccounts 
47services
48statefulsets.apps
49storageclasses.storage.k8s.io
50validatingwebhookconfigurations.admissionregistration.k8s.io
51volumeattachments.storage.k8s.io

10.Editing resources

Edit a service

1kubectl edit service <service>

Edit a service with your favorite text editor

1KUBE_EDITOR="vim" edit service <service>

Note: Change service by any editable resource type like pods.

11.Deleting Resources

Delete a resource using the type and name specified in 

1kubectl delete -f <file>

Delete pods and services with same names

1kubectl delete pod,service <name1> <name2>

Delete pods and services with a custom label

1kubectl delete pods,services -l <label-name>=<label-value>

Delete all pods and services in a namespace

1kubectl -n <namespace> delete pods,services --all

Delete all resources in a namespace

1kubectl delte <namespace>

12.All get commands

1kubectl get all
2kubectl get pods
3kubectl get replicasets
4kubectl get services
5kubectl get nodes
6kubectl get namespaces
7kubectl get configmaps
8kubectl get endpoints

13.Abbreviations / Short forms of resource types

Resource type Abbreviations
componentstatuses cs
configmaps cm
daemonsets ds
deployments deploy
endpoints ep
event ev
horizontalpodautoscalers hpa
ingresses ing
limitranges limits
namespaces ns
nodes no
persistentvolumeclaims pvc
persistentvolumes pv
pods po
podsecuritypolicies psp
replicasets rs
replicationcontrollers rc
resourcequotas quota
serviceaccount sa
services svc

14.Verbose Kubectl

1kubectl run nginx  --image=nginx --v=5
Verbosity Description
--v=0 Generally useful for this to always be visible to a cluster operator.
--v=1 A reasonable default log level if you don't want verbosity.
--v=2 Useful steady state information about the service and important log messages that may correlate to significant changes in the system. This is the recommended default log level for most systems.
--v=3 Extended information about changes.
--v=4 Debug level verbosity.
--v=6 Display requested resources.
--v=7 Display HTTP request headers.
--v=8 Display HTTP request contents.
--v=9 Display HTTP request contents without truncation of contents.

(Table source: K8s docs)

15.Cluster

Display addresses of the master and services

1kubectl cluster-info

Dump cluster state to STDOUT

1kubectl cluster-info dump

Dump cluster state to a file

1kubectl cluster-info dump --output-directory=</file/path>

Compares the current cluster state against the state that the cluster would be in if the manifest was applied

1kubectl diff -f ./my-manifest.yaml

List all images running in a cluster

1kubectl get pods -A -o=custom-columns='DATA:spec.containers[*].image'

16.Kubectl context

Show merged kubeconfig settings

1kubectl config view

Use multiple kubeconfig

1KUBECONFIG=~/.kube/config1:~/.kube/config2:~/.kube/config3

Get a list of users

1kubectl config view -o jsonpath='{.users[*].name}'

Display the first user

1kubectl config view -o jsonpath='{.users[].name}'

Get the password for the "admin" user

1kubectl config view -o jsonpath='{.users[?(@.name == "admin")].user.password}'

Display the current context

1kubectl config current-context

Display list of contexts

1kubectl config get-contexts

Set the default context to 

1kubectl config use-context <cluster>

Sets a user entry in kubeconfig

1kubectl config set-credentials <username> [options]

Sets a user with a client key

1kubectl config set-credentials <user> --client-key=~/.kube/admin.key

Sets a user with basic auth

1kubectl config set-credentials --username=<username> --password=<password>

Sets a user with client certificate

1kubectl config set-credentials <user> --client-certificate=<path/to/cert> --
2embed-certs=true

Set a context utilizing a specific config file

1kubectl config --kubeconfig=<config/path> use-context <cluster>

Set a context utilizing a specific username and namespace.

1kubectl config set-context gce --user=cluster-admin --namespace=foo \
2  && kubectl config use-context gce

17.Alias

Create an alias on *nix

1alias k=kubectl

Create an alias on Windows

1Set-Alias -Name k -Value kubectl

18.Kubectl imperative (create) vs declarative (apply)

Create

You tell your cluster what you want to create, replace or delete, not how you want you it to look like.

1kubectl create -f <filename|url>
2kubectl delete deployment <deployment-name>
3kubectl delete deployment <deployment-filename>
4kubectl delete deployment <deployment-url>

Apply

You tell your cluster how you want it to look like. The creation, deletion and modification of objects is done via a single command. The declarative approach is a statement of the desired end result.

1kubectl apply -f <filename|url>
2kubectl delete -f <deployment-filename>
3kubectl apply -f <deployment-filename>

If the deployment is deleted in , it will also be deleted from the cluster.

Posts in this Series