Microsoft Fleet Intelligence
Monitor and analyze deployed robot fleet telemetry — IoT Operations data collection, Grafana dashboards, drift detection, and Fabric Real-Time Intelligence analytics.
Prerequisites
| Requirement | Purpose |
|---|---|
| Azure IoT Operations | Edge telemetry collection |
| Azure Event Hubs | Cloud telemetry ingestion |
| Grafana | Fleet dashboards and alerting |
| Microsoft Fabric | Real-Time Intelligence KQL analytics |
az CLI |
Azure authentication |
kubectl |
Cluster access for IoT Operations |
Setup Workflow
Deploy fleet intelligence components in order:
Step 1 — Deploy IoT Operations
fleet-intelligence/setup/deploy-iot-operations.sh
Step 2 — Deploy telemetry pipeline
fleet-intelligence/setup/deploy-telemetry-pipeline.sh
Step 3 — Deploy dashboards
fleet-intelligence/setup/deploy-dashboards.sh
Step 4 — Deploy Fabric Real-Time Intelligence
fleet-intelligence/setup/deploy-fabric-rti.sh
Telemetry Schemas
| Schema | Path | Description |
|---|---|---|
| Policy Execution | telemetry/schemas/policy-execution.schema.json |
Inference metrics and action outputs |
| Robot Health | telemetry/schemas/robot-health.schema.json |
Hardware status and connectivity |
Drift Detection
Drift detection monitors deployed policies for performance degradation. When drift exceeds configured thresholds, retraining triggers initiate automated training pipelines.
| Component | Path | Description |
|---|---|---|
| Detection | drift/detection/ |
Statistical tests against training baselines |
| Alerting | drift/alerting/ |
Threshold evaluation and notification routing |
| Triggers | drift/triggers/ |
Automated retraining pipeline launching |
Key Files
| File | Description |
|---|---|
fleet-intelligence/README.md |
Domain overview |
fleet-intelligence/telemetry/routing/edge-to-eventhub.yaml |
Edge-to-cloud routing config |
fleet-intelligence/dashboards/grafana/fleet-overview.json |
Grafana dashboard definition |
fleet-intelligence/dashboards/fabric/fleet-kql-queries.kql |
Fabric KQL analytics queries |