The demo-ai-gitops-catalog repository is a catalog of configurations designed to provision infrastructure on OpenShift that supports machine learning and artificial intelligence workloads. Written primarily in HTML, the project serves as a practical resource for organizations looking to deploy AI and ML workloads on OpenShift using GitOps principles. The repository is maintained by Red Hat's North America Solutions and Sales Acceleration team and is published at redhat-na-ssa.github.io/demo-ai-gitops-catalog/.
The core purpose of this repository is to provide a curated collection of configurations, demonstrations, and workshops that help teams implement GitOps workflows for AI and ML infrastructure on OpenShift. Rather than focusing solely on operator installation like the related GitOps Catalog project, this repository goes deeper by offering kustomized cluster configurations and ArgoCD application definitions that address the specific needs of machine learning and artificial intelligence deployments. Users are directed to explore the kustomized configs in the components/cluster-configs directory and the ArgoCD apps in components/argocd/apps for practical examples and ideas.
The repository requires OpenShift 4.14 or later and specifies different role requirements depending on the scope of deployment. Cluster-level configurations require cluster-admin privileges, while namespaced components only need self-provisioner roles. The README provides tested deployment options through the Red Hat Demo Platform, including AWS with OpenShift Open Environment and One Node OpenShift configurations, with specific node size recommendations for optimal performance.
Getting started with the catalog involves installing the OpenShift Web Terminal operator and then selecting from available demos to deploy. The repository supports multiple deployment approaches, including direct use of the OpenShift web console terminal, local shell environments, and an automated bootstrap script that handles configurations requiring imperative steps or user interaction. Users can cherry-pick individual kustomized app configs and cluster configs rather than deploying everything at once, with operator installations available through simple oc apply commands.
The repository includes a scripts library containing common operational functions available both in local bash shells and in the enhanced web terminal. The README documents a known issue where oc apply -k commands may fail on first execution due to how Kubernetes handles custom resources and custom resource definitions, recommending users re-run commands until successful. An apply_firmly function is provided as a workaround that automatically retries operations.
GitGenius activity data shows the repository has minimal but focused engagement, with codekow as the most active contributor tracked across four events, followed by brady-spiva and hemang-rh with one event each. Response latency metrics across two tracked items show a median of zero hours with a mean of 1048.6 hours, indicating variable response times. The repository shares contributors with the longhorn/longhorn project, suggesting some overlap in the development community.
The project is classified across multiple domains including AI integration, GitOps workflows, Kubernetes automation, cloud-native deployment, catalog management, continuous delivery, DevOps practices, machine learning, infrastructure as code, and Red Hat technologies. The README explicitly warns that the repository is subject to frequent breaking changes and recommends always referencing specific commit hashes or tags rather than relying on the main branch for stability.
