Kubernetes Concepts

To set up the Kubernetes (K8s) infrastructure needed to integrate with Anzo, you use scripts that are supplied by Cambridge Semantics and the API for your preferred cloud service provider (CSP) to deploy a K8s cluster. The cluster includes a K8s API server, which manages all communication for the cluster.

In the cluster, you create a number of node pools or node groups. A node pool or node group is a group of nodes within a cluster that all have the same configuration. Different node pools are designed based on machine types and specific properties to be set on each node. The nodes are tuned to host a particular type of pod. A pod is an instance of an application, i.e., a container of images. The diagram below shows a high level view of a K8s cluster:

For example, an AnzoGraph node pool contains the type of nodes that are suitable for running pods with AnzoGraph images.

Node pools can be configured so that they are static or autoscaling. In static node pools, the nodes are deployed in the K8s cluster and remain provisioned even if they do not run an application. If a node pool is configured with an autoscaler, nodes are not deployed unless resources are requested. When the resources are no longer in use, the autoscaler deprovisions the nodes.

For more information about node pools and other requirements, see Anzo K8s Requirements.