This article contains general instructions for installing the Automation Suite in a Kubernetes cluster.
Automation Suite supports two primary modes of installation, as shown in the following table.
This installs Automation Suite on one machine. Recommended for evaluation and demo purposes. We do not support High Availability (HA) in this mode.
For preparing a single-machine setup, see Single-node machine requirements.
This installs Automation Suite on multiple machines. Recommended option for production deployments. The default configuration and recommended option is to enable HA, but you can override this and disable it.
For preparing a setup for multiple machines, see Multi-node machine requirements.
A Kubernetes cluster is a set of nodes that run containerized applications.
A node refers to any machine (bare metal, virtual machine, etc.).
A server node is defined as a machine (bare-metal or virtual) running the cluster management server (i.e., Rancher).
An agent node is defined as a machine running the worker pods (the functional services). A machine can be designated to be used as both server and agent. Having separate server and agent nodes in a deployment is a topology design decision.
Air-gapped environment refers to a set-up where the machines do not have access to internet.
Ensure you have the infrastructure for running the cluster setup and you’ve completed the steps in the prerequisites documentation before beginning the installation process.
If you are installing on Azure, follow the guide below to configure your resources:
Sample Azure infrastructure setup.
Make sure to read the in-depth prerequisite guide to set up your infrastructure based on the deployment you will install, Use this to evaluate what hardware or cloud resources you need to procure or budget for.
Hardware and software requirements.
We support additional add-ons, primarily for enabling AI workloads:
Configuring the GPU on any agent node for more on adding a GPU to the cluster.
Installing Task Mining for more on adding a dedicated node for running background jobs in Task Mining.
AI Center provides models as ML Packages, which are distributed as
.zip archives uploaded to AI Center’s storage. When a user deploys an ML Skill or a Training Pipeline, these ML Packages are deployed onto a base image. You can download an On-Demand bundle which comprises of AI Center, DU and TM related images separately.
See the following installation guides:
Updated a day ago