Triton Control Documentation

Triton Control is a web application for managing and operating NVIDIA Triton Inference Server environments. The primary deployment target is Kubernetes through the Helm chart in charts/triton-control.

Core capabilities:

• existing Triton instance registration and management
• self-deployed Triton serving workflows when Triton Control runs in Kubernetes
• per-user browser-based Development workspaces backed by code-server
• user management and instance access control
• model inference workflows with model configuration inspection
• S3-backed model repository integration with reusable S3 profiles and an integrated S3 Browser
• Perf Analyzer workflows when Triton Control runs in Kubernetes
• Kubernetes-managed MLflow tracking with persistent storage and an embedded, authenticated MLflow UI
• embedded Argo Workflows UI and API through an authenticated backend proxy

The same application can also run with Docker Compose or Podman Compose for local evaluation, and with separate Python/npm processes for development.

What This Documentation Covers

• Getting Started: Kubernetes-first setup plus Docker, Podman, and local development paths.

Product:

• User Guide: user workflows for dashboard, instances, inference, profile, S3, Add Deployment, Perf Analyzer, MLflow, and Add Instance.
• User Management: roles, local users, OIDC users, approvals, and instance assignment.
• Development Workspaces: the Kubernetes-backed, browser-based code-server workspace and its Triton deployment extension.
• Argo Workflows: installation, authenticated proxy, Kubernetes layout, pod security, RBAC, and Workflow credentials.

Architecture:

• Architecture Overview: high-level architecture, runtime boundaries, and core system interactions.
• Architecture Backend Components: backend service components and responsibilities.
• Model Config Validation: model repository validation behavior and configuration rules.

Operations:

• Configuration: backend environment variables, OIDC source modes, and GitOps examples.
• Deployment: Kubernetes prerequisites, Helm deployment, Docker Compose, and Podman Compose.
• Security: secrets, OIDC, TLS, S3 credentials, and proxy headers.
• Troubleshooting: common Docker, database, and Triton issues.

Reference:

• API: OpenAPI, Swagger, ReDoc, and generated frontend client notes.

Project:

• Roadmap: planned product changes for upcoming versions.
• Local Project Development: local Python/npm workflow, tests, and CI checks for Triton Control itself.

Main Components

• triton-frontend/: Angular Material frontend.
• triton-backend/: Python FastAPI backend.
• charts/triton-control/: Helm chart.
• Dockerfile: combined frontend/backend/Nginx image.
• compose.yaml: Docker Compose stack.
• podman-compose.yaml: Podman Compose stack.