Skip to main content
In this lesson, we explore the concept of namespaces in Kubernetes and how they help organize and isolate resources within your cluster.

​
Understanding Namespaces Through Analogy

Imagine there are two boys named Mark. To differentiate between them, you refer to them by their last names—Smith and Williams. They come from different houses where people often use first names for those familiar with them. However, when addressing someone from another house or an outsider, the full name is used. In Kubernetes, these “houses” represent namespaces. They allow you to group and manage resources differently based on their context and intended use.
The image shows two houses labeled "Mark Smith" and "Mark Williams," with speech bubbles indicating confusion between the two names.

​
Default Namespace and System Namespaces

By default, when you create objects such as pods, deployments, and services in your cluster, they are placed within a specific namespace (similar to being “inside a house”). The default namespace is automatically created during the Kubernetes cluster setup. Additionally, several system namespaces are created at startup:
  • kube-system: Contains core system components like network services and DNS, segregated from user operations to prevent accidental changes.
  • kube-public: Intended for resources that need to be publicly accessible across users.
If you’re running a small environment or a personal cluster for learning, you might predominantly use the default namespace. In enterprise or production environments, however, namespaces provide essential isolation and resource management by allowing environments like development and production to coexist on the same cluster.
The image illustrates Kubernetes namespace isolation with three houses labeled "kube-system," "Default," and "kube-public," each containing a circle, triangle, and square.

​
Isolating Resources with Namespaces

Namespaces allow you to set distinct policies and resource limits for different environments. This isolation prevents one namespace from interfering with another. For instance, you can apply separate resource quotas for CPU, memory, and the total number of pods to ensure fair usage across environments.
The image illustrates Kubernetes namespace resource limits across nodes, showing different environments (Default, Prod, Dev) with associated icons and resource distribution.
Within a single namespace, resources can refer to each other directly via their simple names. For example, a web application pod in the default namespace can access a database service simply by using its service name:
The image depicts a diagram with a house shape labeled "Default," containing a circle, triangle, and square labeled "web-pod," "db-service," and "web-deployment," respectively.
If the web app pod needs to communicate with a service located in a different namespace, you must use its fully qualified DNS name. For example, connecting to a database service named “db-service” in the “dev” namespace follows this format:
Here, “svc” indicates the service subdomain, followed by the namespace (“dev”) and the service name, ending with the default domain “cluster.local”.

​
Managing Namespaces with kubectl

Using kubectl, you can manage resources across different namespaces. Below are some commonly used commands.

​
Listing Pods in a Namespace

To list all pods in the default namespace:
To list pods within the kube-system namespace:

​
Creating Pods in Specific Namespaces

When creating a pod without specifying the namespace, it is placed in the default namespace:
To create the same pod in the “dev” namespace, you can either include the namespace option:
Or define the namespace within the pod definition file:

​
Without Namespace Specification

​
With Namespace Specification

​
Creating a Namespace

You can create a namespace using a YAML file. For example, create a file named namespace-dev.yml:
Then run:
Alternatively, create a namespace directly through the command line:

​
Setting the Default Namespace for Your Context

If you’re working across multiple namespaces and want to avoid repeatedly specifying the namespace flag, you can set the default namespace for your current context:
After setting this, running:
will automatically list pods in the “dev” namespace. To list pods from all namespaces, use:
Contexts are used to manage multiple clusters and user environments within a single configuration. While switching namespaces is simple, managing contexts is a broader topic that warrants further exploration.

​
Controlling Resource Usage with ResourceQuotas

To ensure that no single namespace overconsumes cluster resources, Kubernetes allows you to define ResourceQuotas. For example, create a file named compute-quota.yaml with the following content:
Apply this configuration with:
This configuration guarantees that the “dev” namespace does not exceed the specified resource limits.

​
Conclusion

Namespaces are a fundamental component in Kubernetes, enabling you to segment and manage resources effectively. Whether you’re isolating system components or separating development and production environments, using namespaces along with appropriate policies and resource quotas leads to a more efficient and organized cluster management. Practice these concepts and explore additional Kubernetes functionalities to deepen your understanding. Happy clustering! For further reading, check out these resources:

Watch Video

Practice Lab