Stack Introduction
Who is this documentation for?
This section of the documentation is designed for infrastructure teams, system administrators, and DevOps engineers responsible for:
- Installing and configuring the ConfidentialMind stack
- Managing the underlying infrastructure
- Maintaining the stack components
- Configuring security and access controls
- Managing container registries and deployments
If you're a developer integrating AI into your products on top of an existing ConfidentialMind installation, see Getting Started and the AI APIs instead.
Getting Started
For complete installation instructions, visit our Installation Guide.
What is the ConfidentialMind Stack?
The ConfidentialMind stack is a Kubernetes-based containerized set of services and infrastructure required to run LLMs and LLM-enabled applications on any given cloud or private environment. The stack and included installation tools abstract the deployment and management of GenAI applications to a level where a deep understanding of Kubernetes or AI-ops is not required for application developers.
Key features:
- Complete Kubernetes-based infrastructure for AI applications
- Built-in security and access management
- Automated deployment and scaling
- Integrated monitoring and logging
- Support for both cloud and on-premises deployments
Core Components
The ConfidentialMind stack infrastructure is built on two main Helm charts:
-
stack-base: The primary Helm chart that deploys core services:
- Admin portal
- Model deployment services
- Database management
- API gateway
- Authentication services
-
web-service-chart: Used by the CI/CD system to deploy individual services:
- Defines Kubernetes resources for each service
- Manages service-specific configurations
- Handles service dependencies
For a detailed breakdown of all components, see our Platform Components documentation.
Configuration and Customization
Container Registry
The stack comes pre-configured with the confidentialmind.azurecr.io registry, which provides:
- Ready-to-use components
- Template applications
- Helm charts
Your license includes read-only credentials for this registry.
Available Services
The ConfidentialMind stack includes various pre-configured services:
- Template applications
- AI model deployments
- Database services
- API connectors
Learn more about available services and adding custom ones in our Default Services documentation.
Next Steps
- Review the Installation Guide for detailed setup instructions
- Understand the Platform Components for better system management
- Learn about Default Services and custom service deployment
- Review Graceful Node Shutdown for node and host maintenance practices
Remember: While the ConfidentialMind stack simplifies many aspects of running AI infrastructure, having a basic understanding of Kubernetes and container orchestration will be helpful for successful deployment and management.