ML Provisioner Tool Documentation¶

Welcome to the ML Provisioner Tool documentation. This tool provisions and manages AWS ML product infrastructure using tier-based CloudFormation templates (starter, professional, enterprise).

📚 Getting Started (Read First)¶

Start here if you’re new to the ML Provisioner.

README: Quick start and overview - What the tool does, common commands, tier-based architecture, and AWS credentials setup.
USER_GUIDE: Complete command reference - All 12 actions (validate-config, create-prov-template, deploy-product, delete-product, etc.), deployment workflows, volume mounts, and best practices.

📖 Core Documentation (Essential)¶

CONFIGURATION: Configuration reference - Config file naming convention, discriminators (source_control, vpc_mode, vpc_source), and all enterprise tier combinations.
CONFIGURATION_GUIDE: Configuration guide - Step-by-step decision tree for selecting the correct config file, verifying prerequisites, and populating config values before deployment.
NAMING_CONVENTIONS: Naming conventions - Resource naming patterns, ml_name construction, artifact filename conventions, and SSM parameter paths.
PREREQUISITES: Pre-deployment requirements - S3 source control bucket setup, VPC and route table requirements for enterprise tier, standalone vs bundle networking.

🔒 Security¶

IAM_PERMISSIONS: Required AWS permissions - Least-privilege IAM policy for CloudFormation, SageMaker, CodePipeline, CodeBuild, KMS, VPC, and SSM operations.
SECURITY_GUIDELINES: Security guidelines - KMS encryption, permission boundaries, VPC endpoint security groups, and compliance monitoring.

🔧 Operations (Day-to-Day Use)¶

UPDATE_PROCEDURES: Update and maintenance procedures - How to update the Docker image, modify tier configurations, and perform maintenance tasks.

🐛 Troubleshooting¶

TROUBLESHOOTING: Common issues and solutions - AWS credentials, configuration errors, CloudFormation errors, VPC integration issues, and Docker errors.

📊 Advanced Topics (Optional)¶

RESOURCES_EXPLAINED: AWS resources explained — Educational guide to every AWS resource provisioned by the ML Provisioner, what it does, why it exists, and how it relates to other resources in the stack.
APPLICATION_ARCHITECTURE: Complete architecture documentation - Tier-based design, CFN generator architecture, blueprint system, SSM integration, and deployment patterns.
INTEGRATION_EXAMPLES: Integration examples - How to consume ML stack outputs from SSM Parameter Store in SageMaker, CI/CD pipelines, and downstream provisioners.
MIGRATION_GUIDE: Migration guide - Migrating from manually created ML infrastructure or other IaC tools to the ML Provisioner.

📋 Reference¶

ROADMAP: Future features and plans - Planned features, enhancement requests, and product roadmap.
RELEASE_NOTES: Version history and changes - Release history, new features, bug fixes, and upgrade notes.
SAMPLE_REPORTS: Sample HTML reports - Example pre-deployment review report and deployment report.

📚 Getting Started:

🚀 Onboarding Examples:

Onboarding Examples

📖 Core Documentation:

🔒 Security & Compliance:

🐛 Troubleshooting and Support:

📊 Advanced Topics:

📋 Reference:

📖 Recommended Reading Order¶

For New Users¶

README - Overview and quick start
User Guide - Complete command reference
Configuration Reference - Configuration parameters
Prerequisites - Pre-deployment requirements
Troubleshooting - Common issues

For Production Deployment¶

Prerequisites - Ensure all prerequisites are met
Configuration Reference - Production configuration
Security Guidelines - Security hardening

🏗️ ML Product Tier Architecture¶

The ML Provisioner supports three tiers of ML product infrastructure:

Starter: SageMaker Project, Model Registry, CodeCommit/S3 repositories, CodeBuild, CodePipeline, IAM roles.
Professional: All Starter resources plus S3 artifacts bucket, EventBridge rule and pipe for automated deployment, CloudWatch dashboard, and enhanced managed policies.
Enterprise: All Professional resources plus KMS encryption, compliance log group, CloudWatch alarms, SNS alerting, VPC endpoints, and permission boundaries.

📝 Documentation Conventions¶

UPPERCASE_WITH_UNDERSCORES.md - Main documentation files
All paths are relative to the docs/ directory
Code examples use bash syntax unless otherwise noted
AWS resource names follow the pattern: {company_prefix}-{env}-{tenant_id}-{region}-{use_case}-ml

🔄 Documentation Updates¶

This documentation is version-controlled and updated with each release. See Release Notes for documentation changes in each version.