Model as a Service for Red Hat Demo Platform
No results found.
LiteMaaS is a Model-as-a-Service platform that provides Red Hat Demo Platform users with managed access to AI/LLM models. It handles API key management, model subscriptions, usage tracking, rate limiting, and budget controls — so workshop attendees and developers get a simple OpenAI-compatible API endpoint without managing any infrastructure.
This documentation covers the RHDP deployment using the rhpds.litemaas Ansible collection. For the upstream LiteMaaS application, see rh-aiservices-bu/litemaas.
Full component diagram — all LiteMaaS layers, PostgreSQL, Redis, model servers, GitOps flow, and sequence diagrams. Click any diagram to expand.
What LiteMaaS is, the problem it solves, key features, role hierarchy, and the end-to-end user flow.
Model capability types (Chat, Embeddings, Document Conversion), API endpoints, and curl/SDK examples.
Prerequisites, Ansible collection variables reference, step-by-step deployment, post-install checklist.
Upgrading, scaling, key cleanup cronjob, adding models, LITELLM_AUTO_SYNC, DB sync, admin operations.
Grafana dashboards, vLLM and GPU metrics, ServiceMonitors, LiteMaaS usage analytics, and the automated key cleanup cronjob.
10 common issues with exact fix commands — model sync, OAuth, migrations, cache, OOMKills, timeouts.
rhpds/rhpds.litemaas on GitHub — Ansible roles, playbooks, templates and scripts for RHDP deployment.