Skip to main content

Overview

Prizm features a cloud-native, containerized architecture that enables:
  • Seamless integration with existing data ecosystems
  • Horizontal scalability to handle enterprise-level data volumes
  • Deployment flexibility — cloud, on-premises, or hybrid
  • Open APIs for custom extensions and integrations

High-Level Architecture


Core Architectural Components

AI-Native Control Plane

The control plane is the brain of the Prizm platform:
ComponentResponsibility
Agent OrchestratorPlans, routes, and coordinates agent tasks with memory of prior context
Event BusDecoupled message passing between agents and platform services
Prioritized Task QueueRisk-based scheduling ensures critical assets are always processed first
Policy EngineEnforces SLAs, guardrails, RBAC, and ABAC rules across all operations

Knowledge & State Store

Prizm maintains a rich, interconnected knowledge base:
StoreContents
Unified CatalogAll assets — tables, models, pipelines, dashboards — with owners and tags
Semantic LayerBusiness terms, KPIs, glossaries, and domain context
Knowledge GraphRelationships, lineage, and dependency maps
Vector StoreRunbooks, incident histories, and documentation for RAG-based AI
Operational LogSingle unified observability table across all agents

Product Surface

  • Prizm APIs — GraphQL and REST APIs for programmatic access
  • Prizm UI — Full-featured web interface for catalog, metrics, issues, and governance
  • Alerting & Integrations — Slack, Email, Jira, PagerDuty, ServiceNow

Sub-System Documentation

Data Model

Core entity relationships and schema design.

Metrics

Metric types, execution engine, and scoring.

Score Entity

Quality score schema, attributes, and relationships.

Scheduling

CRON, event-based, and intelligent schedule management.

Notifications

Alerting channels, routing, and notification policies.

Infrastructure

Deployment topology and infrastructure requirements.