R
RobotOps
Technical Architecture

Edge-first observability, not another control system

RobotOps runs as a lightweight observability layer alongside your robots. It ingests telemetry from vendor APIs and message buses, normalizes it into a unified schema, and surfaces insights—without ever touching robot control loops.

Data Sources
🤖

Robot APIs

Locus, 6 River, MiR, Fetch vendor APIs

📡

Message Buses

MQTT, Kafka, VDA 5050 topics

📦

WMS / WES

Order, mission, and zone data

→
Edge Gateway
⚡

RobotOps Edge

Normalize telemetry, aggregate metrics, buffer data, detect anomalies locally

→
Cloud Platform
📊

Analytics

Fleet-wide metrics, predictions, baselines

🔔

Alerts

Smart routing to Slack, Teams, email

🖥️

Dashboards

Real-time fleet visibility and AI Copilot

⚡ Edge Gateway

Runs on-premise on an industrial PC or VM in the same VLAN as your robots. Read-only connections to vendor APIs and message buses.

✓ Connects to vendor APIs and webhooks for robot state
✓ Subscribes to MQTT/Kafka for real-time events
✓ Performs lightweight aggregation and anomaly flags
✓ Buffers locally during network issues

☁️ Cloud Platform

Receives normalized signals from edge gateways and provides fleet-wide analytics, predictions, and dashboards.

✓ Multi-site, multi-vendor time series storage
✓ ML-based baseline learning and predictions
✓ Health scores (MTBF, MTTR, fault rates)
✓ API and webhook exports

🛡️ Safety & Non-Control

RobotOps is observability-first. It never sends commands to robots or intervenes in safety loops.

✓ Does NOT send commands to robots
✓ Does NOT require routing mission traffic
✓ Connects only to telemetry interfaces
✓ Keeps control logic and PLCs out of scope

🔌 Integration Patterns

RobotOps fits into your existing infrastructure without replacing orchestration or WMS systems.

✓ HTTP/REST APIs for OEM robot data
✓ MQTT/Kafka subscriptions for events
✓ WMS/WES bridges for order correlation
✓ Custom adapters via Python SDK
Data Model

Vendor-neutral telemetry schema

RobotOps normalizes signals from any vendor into a unified model for fleet-wide observability.

🤖 Robots

  • Identity, vendor, model
  • Site and zone assignment
  • Capabilities and lifecycle state
  • Current status and location

📋 Missions

  • Task type and priority
  • Start/end timestamps
  • Associated orders
  • Completion status

💚 Health

  • Fault codes and errors
  • Aborted missions
  • MTBF / MTTR metrics
  • Component health scores

📈 Performance

  • Missions per hour
  • Queue and wait times
  • Productive vs idle time
  • Throughput by zone

🔋 Telemetry

  • Battery state of charge
  • Motor temperatures
  • Velocity and position
  • Sensor readings

🚨 Incidents

  • Derived alert objects
  • Severity and priority
  • Recommended actions
  • Assignment and status

Questions about architecture?

Our team can walk through how RobotOps fits into your specific infrastructure.

Talk to an engineer