Skip to main content

Namespaces

Namespaces Part 1
Namespaces Part 2

In Kubernetes namespaces allow a separation of Kubernetes objects in a cluster.

By default, all operations are executed in the default namespace.

While namespaces can be used to separate workloads of different users such as different development teams, the level of isolation provided by namespaces is not sufficient for tenant isolation. So in a multi-user environment where there is no trust relationship among the users, the deployment of a dedicated Kubernetes cluster per user or user group is common.

The practical usage of namespaces is often to isolate different workloads running within a cluster or separate different development environments such as staging and testing.

Creating a Namespace​

There are multiple ways to interact with Kubernetes and create a namespace. One is to use the kubectl command:

Create your own namespace using kubectl

kubectl create namespace k8s-training

Alternatively, a YAML file can be created, e.g. 30-namespace.yaml:

apiVersion: v1
kind: Namespace
metadata:
  name: k8s-training

and applied using

kubectl apply -f 30-namespace.yaml

Creating a kubectl Context to Work Within a Namespace​

While it is possible to specify namespaces explicitly this would obstruct the portability of a workload description. Hence, it is recommended to avoid specifying namespaces in YAML files if possible.

In order to avoid the necessity to specify a --namespace k8s-training for each kubectl command, it is handy to create a so called context.

kubectl config set-context k8s-training --namespace k8s-training --cluster=minikube --user minikube

Note: Your cluster may have a different cluster name and username. You can obtain both pieces of information using the following command:

kubectl config view

Once the context has been created we can switch to it:

kubectl config use-context k8s-training

From now on all kubectl commands will be executed within the k8s-training namespaces as specified in the k8s-training context.