What is UnderStack?

UnderStack is a collection of open-source tools and technologies that provides flexible, scalable, and cost-effective infrastructure management solution. The focus is on deploying bare metal in data centers with functional DCIM and IPAM.

About the Guide

While the deployment of UnderStack leans heavily into GitOps, it is not meant to be a definitive guide to GitOps. It is also not the only way GitOps can be used with UnderStack but instead focuses on one example deployment installation. It will make a few assumptions and some opinionated choices that may not align with a production best practices installation. Improvements are always welcome.

System Division

A fully deployed UnderStack is divided into three distinct parts, or clusters, that are referred to in the documentation as:

• Management
• Global
• Site(s)

flowchart TD

  A[Management] --> B[Global];
  B[Global] --> C[Site A];
  B[Global] --> D[Site B...];
  B[Global] --> E[Site N];

A fully functioning system only needs one Management cluster, one Global cluster and one or more Site cluster(s). In this configuration, the Management cluster is responsible for utilizing our GitOps tool, ArgoCD to deploy the expected state to all other clusters. While the Global cluster is responsible for hosting any services that are expected to exist only once for a whole system deployment such as the DCIM/IPAM tool. While the Site clusters will run the tools and services that need to live close to the actual hardware.

In fact, one Management cluster can control multiple Global clusters and their associated Site clusters. We call the grouping of the Global cluster and it's associated Site clusters a partition. An example would be a staging partition and a production partition.

flowchart TD

  A[Management] --> |staging| B[Global];
  B[Global] --> C[Site A];
  B[Global] --> D[Site B...];
  B[Global] --> E[Site N];
  A[Management] --> |production| F[Global];
  F[Global] --> G[Site D];
  F[Global] --> H[Site E...];
  F[Global] --> I[Site Z];