Kubernetes is a powerful orchestration platform that can be used to manage large scale deployments of applications. However, managing Kubernetes can be complex and time-consuming, especially if you are not familiar with the command line interface (CLI). To make managing Kubernetes easier, Kubectl provides context-based management. This means that you can use one command to manage all of your nodes in a cluster, regardless of their location or state. In this article, we will show you how to use contexts to simplify your Kubernetes management. We will also provide some tips for using contexts effectively. ..
Kubectl contexts are a mechanism for quickly switching between different clusters, users, and namespaces within the CLI. They make it easier to move between multiple environments without changing your active Kubectl config file.
In this article, we’ll show how you can use Kubectl to create, manage, and select different contexts. Make sure you’ve got Kubectl installed before you continue.
What’s A Context?
Contexts encapsulate the collection of settings that permit a successful connection to a Kubernetes cluster. A context can include the cluster’s URL, a set of user credentials, and the namespace to target by default.
In the absence of contexts, unique Kubernetes environments are often handled by creating a separate config file for each one. You then use the –kubeconfig flag or KUBECONFIG environment variable to load the correct file each time you use Kubectl:
Contexts let you condense details of all your environments into one config file. You can use the default .kube/config for each of your clusters, eliminating CLI flags and environment variables. Kubectl includes commands to switch its active context between the options you’ve created.
Preparing for Contexts
Contexts are managed using the kubectl config command group. As with everything else in Kubectl, your available context list will be loaded from and saved to your active config file. This is determined by KUBECONFIG, –kubeconfig, or the default .kube/config.
To begin using contexts, you need to add a few clusters and credentials to your config file. You can use other kubectl config commands to set these up:
Now your config file contains connection details for two separate Kubernetes clusters. It also holds two pairs of credentials. Next we’ll create a context to link the clusters to their respective credentials. You’ll then be able to use a Kubectl command to jump between the QA and production environments.
Creating a Context
The kubectl config set-context command adds new contexts to your config file. You must specify a name for your context. Use the command’s flags to reference a previously added cluster and user account.
At this point you can run the kubectl config view command to inspect all the changes that have been made to your config file:
The context definitions point to other kinds of object defined elsewhere in the config file.
Selecting and Using Contexts
Contexts are selected with the kubectl context use-context command:
The active context’s name is stored as the value of the current-context field in your Kubectl config file. All Kubectl commands will target the cluster referenced by the selected context.
The ability to quickly change the target environment within Kubectl helps you move between clusters without being overwhelmed by config flags.
Because the selected context persists until another one is selected, you should check Kubectl is targeting the environment you expect before running destructive commands. Use the current-context command to see the selected context’s name:
You can view all the contexts in your currently loaded config file with get-contexts:
Including Namespace Information With Contexts
So far we’ve used contexts to select a cluster and user account. Contexts can include namespace information too. When a context has an assigned namespace, Kubectl commands will automatically include the –namespace flag. You can still use –namespace manually to override the namespace set by the context.
There’s no limit to the number of contexts you can have. Clusters may appear in multiple contexts, letting you define separate contexts for each of your important namespaces. This avoids repetition of the –namespace flag as you inspect different resources in your cluster.
Renaming and Deleting Contexts
Rename contexts using the rename-context command:
To delete a context, pass its name to the delete-context command:
The cluster, user, and namespace referenced by a context are changed by repeating the set-context command with the same context name. You can also make modifications by manually editing your Kubectl config file.
Making Context Switching Even Easier
Kubectl’s integrated context management can be sufficient when you switch clusters on a relatively infrequent basis. However, if you’re constantly changing cluster throughout your day, the relatively verbose use-context command can start to feel repetitive itself.
Kubectx is a Kubectl plugin that can make context switches even easier. It shortens use-context and adds a few extra convenience features:
Depending on your workflow, you may want to keeping using multiple Kubectl config files too. You could use shell aliases and default environment variables to set up a customized workflow that automatically selects a config file and context for each new terminal window.
Summary
Kubectl contexts are a way to encapsulate multiple logical cluster connections in a single config file. Each context is assigned a cluster URL, user account, and namespace. Kubectl commands will target the selected context’s cluster using the referenced credentials.
You can set up contexts with kubectl config set-context or by manually editing your .kube/config file. Kubectl also includes commands for managing the named cluster connections and user accounts you can reference in your contexts.
