Table of Contents

Kong, by design, is configured dynamically and responds to the changes in your infrastructure.

Kong is deployed onto Kubernetes along-with a Controller, which is responsible for configuring Kong.
All of Kong’s configuration is done using Kubernetes resources, stored in Kubernetes’ data-store (etcd). You can use the power of kubectl (or any custom tooling around kubectl) to configure Kong and get benefits of all Kubernetes, such as declarative configuration, cloud-provider agnostic deployments, RBAC, reconciliation of desired state, and elastic scalability.

Kong is configured using a combination of Ingress resource and Custom Resource Definitions(CRDs).

You can learn more about the design and deployment model here.

You can use one of the following tools to install Kong on Kubernetes:

YAML manifests

To get started with Kong via kubectl, a simple deployment can be achieved using:

kubectl apply -f

Please note that this does not serve as a production-grade deployment. There are certain knobs that you would want to tweak based on your use-case:

  • Replicas: ensure that you are running multiple instances of Kong to protect against outages from a single node failure.
  • Kong’s performance optimizaiton: You have to tweak various memory settings of Kong as and how you tailor the deployment to your use-case.
  • Load-balancer: ensure that you are running a Layer-4 or TCP based balancer in-front of Kong. This makes it possible for Kong to serve TLS certificate and take advantage of integration with cert-manager.

Helm Chart

Kong has an official Helm chart.

To deploy Kong onto your Kubernetes cluster with Helm:

helm repo update # get the latest charts
helm install stable/kong

Documentation on the Helm chart can be found on Helm’s Hub.


Kong’s manifests for Kubernetes can be declaratively pactched using kustomize.

Here is an example of a remote custom build:

kustomize build

There are kustomizations avaialable in our repository for different types of deployments.

Managed Kubernetes by cloud-provider

To install Kong on a managed Kubernetes offering such as GKE, EKS, AKS, etc, please ensure that you have set up your Kubernetes cluster on the cloud-provider and have kubectl configured on your workstation.

Once you’ve it configured, installation for any cloud-provider remains the same, you can use any of the above methods to install Kong.

Each cloud-provider has some minor variation in how they allow configuring specific resources like Load-balancers, volumes, etc, and we recommend following their documentation to tweak those settings.

Using a database

We recommend running Kong in the in-memory (also known as DB-less) mode inside Kubernetes as all the configuration is stored in Kubernetes control-plane. This setup drastically simplifies Kong’s operations as one doesn’t need to worry about Database provisioning, backup, availability, security, etc.

In case you decide to use a database, it is recommended that you run the database outside Kubernetes. You can use a service like Amazon’s RDS or a similar managed Postgres service from your cloud-provider to automate database operations.

We do not recommend using Kong with Cassandra on Kubernetes deployments, as the features covered by Kong’s use of Cassandra are taken care of by other means in Kubernetes.


Once you have got Kong installed on Kubernetes, please follow one of our guides to start configuring Kong.

All the documentation on how Kong works with Kubernetes can be found at the following location:

For questions and discussion, please visit Kong Nation.

Get Updates!

Follow Up:

Get future release notes emailed to you.

Having Trouble? We're Here to Help!