gitlab-ce/doc/topics/autodevops/multiple_clusters_auto_devops.md

stage	group	info	title
Deploy	Environments	To determine the technical writer assigned to the Stage/Group associated with this page, see https://handbook.gitlab.com/handbook/product/ux/technical-writing/#assignments	Multiple Kubernetes clusters for Auto DevOps

{{< details >}}

• Tier: Free, Premium, Ultimate
• Offering: GitLab.com, GitLab Self-Managed, GitLab Dedicated

{{< /details >}}

When using Auto DevOps, you can deploy different environments to different Kubernetes clusters.

The Deploy Job template used by Auto DevOps defines three environment names:

• review/ (every environment starting with review/)
• staging
• production

These environments are tied to jobs using Auto Deploy, so they must have different deployment domains. You must define separate KUBE_CONTEXT and KUBE_INGRESS_BASE_DOMAIN variables for each of the three environments.

Deploy to different clusters

To deploy your environments to different Kubernetes clusters:

1. Create Kubernetes clusters.
2. Associate the clusters to your project:
   1. Install a GitLab agent on each cluster.
   2. Configure each agent to access your project.
3. Install NGINX Ingress Controller in each cluster. Save the IP address and Kubernetes namespace for the next step.
4. Configure the Auto DevOps CI/CD Pipeline variables
   • Set up a KUBE_CONTEXT variable for each environment. The value must point to the agent of the relevant cluster.
   • Set up a KUBE_INGRESS_BASE_DOMAIN. You must configure the base domain for each environment to point to the Ingress of the relevant cluster.
   • Add a KUBE_NAMESPACE variable with a value of the Kubernetes namespace you want your deployments to target. You can scope the variable to multiple environments.

For deprecated, certificate-based clusters:

1. Go to the project and select Operate > Kubernetes clusters from the left sidebar.
2. Set the environment scope of each cluster.
3. For each cluster, add a domain based on its Ingress IP address.

{{< alert type="note" >}}

Cluster environment scope is not respected when checking for active Kubernetes clusters. For a multi-cluster setup to work with Auto DevOps, you must create a fallback cluster with Cluster environment scope set to *. You can set any of the clusters you've already added as a fallback cluster.

{{< /alert >}}

Example configurations

Cluster name	Cluster environment scope	KUBE_INGRESS_BASE_DOMAIN value	KUBE CONTEXT value	Variable environment scope	Notes
review	review/*	review.example.com	path/to/project:review-agent	review/*	A review cluster that runs all review apps.
staging	staging	staging.example.com	path/to/project:staging-agent	staging	Optional. A staging cluster that runs the deployments of the staging environments. You must enable it first.
production	production	example.com	path/to/project:production-agent	production	A production cluster that runs the production environment deployments. You can use incremental rollouts.

Test your configuration

After completing configuration, test your setup by creating a merge request. Verify whether your application deployed as a Review App in the Kubernetes cluster with the review/* environment scope. Similarly, check the other environments.