Constellation OverwatchConstellation Overwatch
Concepts

Architecture

System architecture overview for Constellation Overwatch

System architecture overview for Constellation Overwatch.

Overview

Constellation Overwatch is designed as a distributed, event-driven C4 Industrial Data Fabric that brings together telemetry, video, and AI processing in a unified platform built on NATS JetStream for reliable, low-latency messaging with atomic operations and durable streams.

System Diagram

Constellation Overwatch Architecture

API Service Diagram

Data Flow Sequence

Core Components

Data Fabric Layer

ComponentTechnologyPurpose
MessagingNATS JetStreamReal-time pub/sub and persistent messaging
DatabaseSQLiteEmbedded relational storage with auto-initialization
APIGo + TemplRESTful API and server-rendered UI
Web DashboardDatastar + SSEReal-time reactive UI with Server-Sent Events
MetricsPrometheusNATS and worker collectors for observability

Workers

Background event processors that drive the event-driven architecture:

WorkerPurpose
EntityWorkerProcesses entity creation, updates, deletions, and status changes
TelemetryWorkerHandles high-frequency sensor data and updates KV state
CommandWorkerRoutes commands to entities and processes acknowledgments
EventWorkerProcesses general events and triggers state sync
VideoWorkerManages video stream lifecycle and WebRTC signaling

NATS Streams

StreamPurpose
CONSTELLATION_ENTITIESEntity lifecycle events (created, updated, deleted, status)
CONSTELLATION_EVENTSGeneral system and integration events
CONSTELLATION_TELEMETRYHigh-frequency sensor and position data
CONSTELLATION_COMMANDSCommand dispatch and acknowledgment

KV Store

BucketPurpose
CONSTELLATION_GLOBAL_STATEReal-time entity state keyed by entity_id, watched by SSE handlers for live UI updates

Edge Layer

ComponentPurpose
Aero Arc RelayMAVLink bridge for UAVs
VisionAI-powered video analysis
FFmpegLow-latency video streaming

Integration Layer

ProtocolUse Case
MAVLinkDrone communication
RTSP/WebRTCVideo streaming
REST/WebSocketAPI and real-time updates

Design Principles

  1. Edge-First - Process data as close to the source as possible
  2. Offline-Capable - Full functionality without internet connectivity
  3. Security-First - Air-gapped deployment support
  4. Lightweight - Minimal resource footprint for embedded systems
  5. Interoperable - Standard protocols and open APIs

On this page