About This Architecture

Azure Unified KPI Metric Collection Platform ingests telemetry from multiple AI agents (IES MINT, Placemat, MSD) via Telemetry SDK and Azure Event Hub, with schema validation and dead-letter handling. Data flows through Microsoft Fabric's three-stage pipeline: raw event ingestion into Delta Tables, ETL transformation via Data Factory and Spark Notebooks with quality checks, and curated KPI aggregation served via Power BI DirectLake and Data Activator alerts. This architecture enforces data contracts, implements RBAC governance, and enables real-time KPI dashboards and automated notifications across enterprise AI scenarios. Fork this diagram to customize agent sources, KPI dimensions, or alert thresholds for your organization's metrics platform. The design demonstrates best practices for scalable, governed data collection at enterprise scale.

People also ask

How do I build a scalable KPI collection platform in Azure that ingests from multiple AI agents, validates schemas, and serves real-time dashboards?

This diagram shows a three-stage Azure architecture: Stage 1 ingests telemetry from IES Agents, Placemat, and MSD agents via Telemetry SDK into Azure Event Hub with schema validation and dead-letter queuing. Stage 2 uses Data Factory and Spark Notebooks to transform raw events into curated KPI Delta Tables with quality checks. Stage 3 serves aggregated KPIs via Power BI DirectLake and triggers ale