This document explains the configuration options available for your Kubernetes
clusters. You can create Kubernetes clusters that function within a single
project, or span across multiple projects, to align with your strategy for
managing container workloads in Google Distributed Cloud (GDC) air-gapped. Cluster
configurations also offer different permissions-based controls for
administrative actions, so GDC can manage your
container workload settings or you can manually configure them based on your
use case for more flexibility.

This document is for the following audience groups:

- IT administrators from the platform administrator group who are responsible
  for creating Kubernetes clusters to host container workloads.

- Application developers from the application operator group who are responsible
  for developing container applications in an air-gapped environment.

For more information, see [Audiences for GDC air-gapped documentation](https://docs.cloud.google.com/distributed-cloud/hosted/docs/latest/gdcag/resources/audiences.md.txt).

## Configuration options

In GDC, there are two cluster configurations that offer
different levels of management and flexibility:

- **Shared cluster**: a multi-project Kubernetes cluster managed by the platform administrator group that offers a full suite of integrated and managed services on the organization level.
- **Standard cluster**: a single-project, self-service Kubernetes cluster managed by the application operator group that offers greater flexibility for custom workloads that might conflict with the managed services in a shared environment.

The following table describes the differences between the shared and standard
Kubernetes clusters:

| Feature | Shared clusters | Standard clusters |
|---|---|---|
| Owner | Platform administrator group | Application operator group |
| Cluster administrator | Platform administrator group | Platform administrator group or application operator group |
| Tenancy | Multiple projects | Single project |
| Lifecycle Management | Create, Read, Update, Delete, and Upgrade | Create, Read, Update, Delete, and Upgrade |
| Monitoring | Prometheus with Grafana Dashboards for Kubernetes | Prometheus with Grafana Dashboards for Kubernetes |
| Ingress and egress across projects | Managed by GDC | Configurable |
| Backup and restore | Cluster, project, and workload | Cluster, project, and workload |
| Kubernetes namespace management | Managed by GDC | Configurable |
| Custom resource and controller | Managed by GDC | Configurable |
| Logging | Managed by GDC | Managed by GDC |
| Audit and Billing | Managed by GDC | Configurable |
| Resource type | Zonal only | Zonal only |
| Surface support | GDC console, API, and Terraform | GDC console, API, and Terraform |
| Managed services | Includes a comprehensive set of protected services that are not configurable. | Includes a minimal set of essential services, providing more service flexibility. |
| Marketplace services | Includes integration with marketplace services. | Not available for integration. |

For more information, see the following cluster configuration sections.

### Shared cluster

The shared cluster provides a system-managed cluster that includes a
set of protected services, such as Istio, Gatekeeper, Managed
Harbor Service, and Nginx. The system-managed approach for your cluster
provides an opinionated Kubernetes cluster configuration that runs in the
`platform` namespace, and can attach to any of your existing projects.

Because the shared cluster can span multiple projects, it is
organization-scoped, which means its availability to some audience groups is
limited. The platform administrator group primarily creates and
manages the shared cluster with very little oversight from application
developers.

Cross-project ingress and egress networking traffic is managed by
GDC. Networking policies are typically handled using a
[project network policy](https://docs.cloud.google.com/distributed-cloud/hosted/docs/latest/gdcag/platform/pa-user/pnp/pnp-overview.md.txt).

The primary use cases for the shared cluster are the following:

- You need a cluster that offers a full ecosystem of managed capabilities by default with little customization required.
- You want to manage your container workloads across multiple projects.
- You have no requirements for migrating workloads from existing cloud environments.

For more information about creating a shared cluster, see
[Create a shared cluster](https://docs.cloud.google.com/distributed-cloud/hosted/docs/latest/gdcag/platform/pa-user/create-user-cluster).

### Standard cluster

The standard cluster provides a configurable Kubernetes cluster that includes a
minimal set of services, such as Prometheus and Grafana. Standard clusters include
only essential services for necessary Kubernetes container workload
functionality, letting you install additional services to customize the
cluster for your use case.

The standard cluster is scoped within a project only, which gives application
developers that are confined within a project direct control over how it
functions. The application operator group primarily creates and
manages the standard cluster with very little oversight from platform
administrators.

In many cases, application developers control cluster networking using standard
Kubernetes networking APIs, instead of relying on
GDC-specific networking configurations. However,
standard clusters can configure a subset of
GDC-specific networking configurations, such as the
following:

- Policies
  - [Intra-project network policies](https://docs.cloud.google.com/distributed-cloud/hosted/docs/latest/gdcag/platform/pa-user/pnp/intra-project-traffic.md.txt#create-standard-cluster-intra)
  - [Cross-project network policies](https://docs.cloud.google.com/distributed-cloud/hosted/docs/latest/gdcag/platform/pa-user/pnp/cross-project-traffic.md.txt#create-standard-cluster-cross)
  - [Organization-external project network policies](https://docs.cloud.google.com/distributed-cloud/hosted/docs/latest/gdcag/platform/pa-user/pnp/organization-external-traffic.md.txt#create-standard-cluster-org-external)
- Load balancing
  - [Internal load balancing](https://docs.cloud.google.com/distributed-cloud/hosted/docs/latest/gdcag/platform/pa-user/ilb-service.md.txt#create-internal-load-balancer-standard)
  - [External load balancing](https://docs.cloud.google.com/distributed-cloud/hosted/docs/latest/gdcag/platform/pa-user/elb-service.md.txt#create-external-load-balancer-standard)
- Cloud NAT
  - [Cloud NAT gateways](https://docs.cloud.google.com/distributed-cloud/hosted/docs/latest/gdcag/platform/pa-user/cloud-nat-standard-clusters.md.txt)

The primary use cases for the standard cluster are the following:

- You need a cluster that offers minimal services by default, which allows for comprehensive configurations to fit your needs.
- You want more control over deeper Kubernetes constructs like custom resource definitions or namespaces.
- You want to manage your container workloads within a single project only.
- You have existing container workloads on an existing cloud environment you want to migrate into your GDC air-gapped environment.

For more information about creating a standard cluster, see
[Create a standard cluster](https://docs.cloud.google.com/distributed-cloud/hosted/docs/latest/gdcag/platform/pa-user/create-standard-cluster.md.txt).

## What's next

- [Deploy a container app](https://docs.cloud.google.com/distributed-cloud/hosted/docs/latest/gdcag/platform-application/deploy-container-workloads.md.txt)
- [High availability for your apps](https://docs.cloud.google.com/distributed-cloud/hosted/docs/latest/gdcag/platform-application/pa-ao-operations/ha-apps/overview.md.txt)
- [Container workloads in GDC](https://docs.cloud.google.com/distributed-cloud/hosted/docs/latest/gdcag/application/ao-user/containers/containers-intro.md.txt)