Skip to content

Glossary#

This glossary is a collection of terms related to Project 2A. It clarifies some of the unique terms and concepts we use or explains more common ones that may need a little clarity in the way we use them.

Beach-head services#

We use the term to refer to those Kubernetes services that need to be installed on a Kubernetes cluster to make it actually useful, for example: an ingress controller, CNI, and/or CSI. While from the perspective of how they are deployed they are no different from other Kubernetes services, we define them as distinct from the apps and services deployed as part of the applications.

Cluster API (CAPI)#

CAPI is a Kubernetes project that provides a declarative way to manage the lifecycle of Kubernetes clusters. It abstracts the underlying infrastructure, allowing users to create, scale, upgrade, and delete clusters using a consistent API. CAPI is extensible via providers that offer infrastructure- specific functionality, such as AWS, Azure, and vSphere.

CAPI provider (see also Infrastructure provider)#

A CAPI provider is a Kubernetes CAPI extension that allows 2A to manage and drive the creation of clusters on a specific infrastructure via API calls.

CAPA#

CAPA stands for Cluster API Provider for AWS.

CAPV#

CAPV stands for Cluster API Provider for vSphere.

CAPZ#

CAPZ stands for Cluster API Provider for Azure.

Cloud Controller Manager#

Cloud Controller Manager (CCM) is a Kubernetes component that embeds logic to manage a specific infrastructure provider.

ClusterIdentity#

ClusterIdentity is a Kubernetes object that references a Secret object containing credentials for a specific infrastructure provider.

Credential#

A Credential is a custom resource (CR) in HMC that supplies 2A with the necessary credentials to manage a specific infrastructure. The credential object references other CRs with infrastructure-specific credentials such as access keys, passwords, certificates, etc. This means that a credential is specific to the CAPI provider that uses it.

Hosted Control Plane (HCP)#

An HCP is a Kubernetes control plane that runs outside of the clusters it manages. Instead of running the control plane components (like the API server, controller manager, and etcd) within the same cluster as the worker nodes, the control plane is hosted on a separate, often centralized, infrastructure. This approach can provide benefits such as easier management, improved security, and better resource utilization, as the control plane can be scaled independently of the worker nodes.

Infrastructure provider (see also CAPI provider)#

An infrastructure provider (aka InfrastructureProvider) is a Kubernetes custom resource (CR) that defines the infrastructure-specific configuration needed for managing Kubernetes clusters. It enables Cluster API (CAPI) to provision and manage clusters on a specific infrastructure platform (e.g., AWS, Azure, VMware, OpenStack, etc.).

Managed cluster#

A Kubernetes cluster created and managed by Project 2A.

Management cluster#

The Kubernetes cluster where 2A is installed and from which all other managed clusters are managed.