Operator Guides

These guides cover deploying, configuring, and integrating Michelangelo AI in a Kubernetes environment. They target platform engineers and infrastructure operators who are responsible for running Michelangelo AI in production and for connecting it to the broader ML infrastructure their teams already use — experiment tracking, model registries, compute clusters, schedulers, and serving frameworks.

Getting Started

For a fresh deployment, follow this recommended reading order:

1. Platform Setup — configure each component (API server, controller manager, worker, UI/Envoy) via ConfigMaps and Kustomize overlays
2. Register a Compute Cluster — connect an existing Kubernetes cluster so Michelangelo AI can dispatch Ray and Spark jobs to it
3. Cluster Setup for Serving — enable model inference on a local or remote cluster
4. Authentication — connect an identity provider and configure RBAC before opening to users

Setup & Configuration

Guide	Description
Helm Chart	Install the Michelangelo AI control plane with Helm — chart layout, values reference, and migration phases
Platform Setup	ConfigMaps and key fields for API server, controller manager, worker, and UI/Envoy
Network & Ingress	Envoy proxy, Ingress setup, TLS with cert-manager, and multi-cluster connectivity
Authentication	OIDC identity provider setup, RBAC, session configuration, multi-tenant isolation
Register a Compute Cluster	Connect an existing Kubernetes cluster to the Michelangelo AI control plane

Platform Components

Guide	Description
Model Registry	Operate Michelangelo AI's built-in model registry, configure storage and RBAC, and integrate with serving and CI/CD
Ingester Controller	Deploy, configure, and operate the ingester that syncs CRDs into MySQL

Jobs & Compute

Guide	Description
Jobs Overview	Ray and Spark job lifecycle, compute selection, and observability
Run a Pipeline on a Compute Cluster	Submit and monitor a Uniflow pipeline on a registered cluster
Extend the Job Scheduler	Custom scheduling backends (Kueue, Volcano) and assignment strategies

Model Serving

Guide	Description
Serving Overview	InferenceServer and Deployment lifecycle, architecture
Cluster Setup for Serving	Configure a cluster for inference
Integrate a Custom Backend	Plugin interfaces for Triton, vLLM, TensorRT-LLM, and custom frameworks

UI

Guide	Description
Deploying the UI	Deploy the Michelangelo AI web UI to Kubernetes
Local UI Development	Run the UI locally for development

Third-Party Integrations

Michelangelo AI is designed to run alongside existing ML infrastructure. The guides below cover making external tools reachable from Michelangelo AI workloads.

Guide	Description
Experiment Tracking Setup	Make an experiment tracking server reachable from task pods — network, ConfigMap injection, auth, and operator/user boundary
Notification Delivery Setup	Wire in email and Slack delivery for PipelineRun state-change notifications — task queue config, activity implementation, and verification
Browse all integrations	MLflow and other third-party integration guides

Operations

Guide	Description
Monitoring & Observability	Prometheus scrape config, key metrics, alerting rules, Grafana dashboards, structured logging
Compliance	SOC 2, GDPR, and HIPAA configuration
Cascade Delete	Pipeline cascade delete (on by default) — propagation policy, RBAC, lifecycle, drain finalizers, per-child safety timeout, and metrics
Troubleshooting	Common failure modes and kubectl diagnostic commands

Architecture & Reference

Guide	Description
API Framework	Architecture overview of the Michelangelo AI API and control plane
SQL Key Concepts and Terms	Metadata schema, table naming, indexed fields, and SQL query patterns