Flowmaster Overview

FlowMaster Microservices Platform Overview

SNAPSHOT: 2026-02-08 11:00 Dubai Time (07:00 UTC) — likely changing soon

Executive Summary

FlowMaster is a microservices-based business process automation platform for designing, executing, and monitoring workflows with AI integration. As of Feb 8 2026, 29 services are deployed and running on a K3S Kubernetes cluster (dev-01 server 65.21.153.235), covering all 79 requirements (R01-R79). Code is deployed but ALL services are UNTESTED — no integration, e2e, or health check testing has been performed.

Source Control: GitLab flow-master group (primary), mirrored to GitHub HCB-Consulting-ME. Code locations: ~/projects/documentation (stale copies), ~/projects/flowmaster/ (working dir, not a git repo).

Deployed Services (29 Running Pods — K3S)

Core Business Logic (13 services)

New Services (6, built Feb 2026)

Rebuilt/Revived Services (3)

Frontend & Companion Apps (7)

Data Intelligence

• Scaled to 0 replicas (replaced by Process Analytics)

Architecture & Data Patterns

Deployment Model (CURRENT)

K3S Kubernetes on dev-01 server (65.21.153.235):

• 29 pods in flowmaster namespace
• Local Docker registry: localhost:30500
• Images deployed via docker save/gzip/scp/docker load (NOT via CI/CD pipelines)
• Code NOT pushed to GitLab repos yet

Network Architecture (Dev-01 Server)

Internet → Nginx (port 80) → K3S ClusterIPs
  / → frontend (10.43.185.219:3000)
  /api/ → api-gateway (10.43.193.134:9000)
  /ws/ → websocket-gateway (10.43.43.253:9010)

WARNING: ClusterIPs are hardcoded in nginx config, will change if services recreated.

Synchronous Communication

HTTP REST via API Gateway (port 9000):

• Frontend → API Gateway → Services
• Service-to-service calls (sync where needed)
• JWT validation at gateway + service level

Asynchronous Communication

Kafka Event Bus:

• Services publish domain events (e.g., task.created, execution.completed)
• Event Bus routes via subscriptions (webhook delivery)

Real-Time Streaming

WebSocket (Socket.io) + Redis:

• WebSocket Gateway (port 9010)
• Services publish to Redis channels
• Gateway broadcasts to connected clients

Databases

• ArangoDB: Process definitions, task graphs (shared container, port 8529, 3 months uptime)
• PostgreSQL: Auth, executions, scheduling, notifications
• Redis: Caching, pub/sub, session management

Key Concepts

Process Explorer vs. Process Designer

• Process Explorer: Full experience - browse processes, view analytics, manage data
• Process Designer: Visual drag-and-drop editor (R77) that sits INSIDE the Explorer

Architectural Decisions (D1-D18)

• D1: SSO KEEP (working on dev-01 server)
• D3: Notification + Communication MERGE
• D4: AI Agent Service = Agent Orchestration
• D5: Learning Management → Agent Service (BUILT)
• D6: Prompt Engineering KEEP SEPARATE (BUILT)
• D7: Internal Data Hub = Knowledge Hub (REVIVED)
• D8: Observability SKIP FOR NOW
• D9: Permissions enhancement SKIP FOR NOW
• D10: SDX is MOST CRITICAL integration
• D11: DXG + Engage REPLACE human task execution UI
• D12: Process Explorer ≠ Process Designer
• D13: SDX REST-only internally; MCP via unified FlowMaster MCP Server
• D14: Agent learning from Engage feedback → Knowledge Hub → Agent Service loop
• D15: Process Analytics replaces empty Data Intelligence
• D16: BAC = process marketplace
• D17: Business Rules = first-class DMN-style objects
• D18: Process Designer = Visio-quality drag-and-drop with swimlanes + inline AI

Deployment Servers

Note: Dev-01 server has ONE FlowMaster environment (not separate dev/test/staging). SDX is part of FlowMaster (currently standalone Docker, should be in K3S). Plane CE is a separate project management tool, not part of the product.

Tech Stack Summary

Backend: 11/13 core services Python/FastAPI, 2 Node.js/TypeScript Frontend: Next.js, React, TypeScript, Radix UI (shadcn/ui), TailwindCSS Source Control: GitLab (primary), GitHub (mirror) Container Orchestration: K3S (demo), Docker Compose (prod/staging/dev) CI/CD: GitLab CI, runner flowmaster-demo-runner v18.8.0

Requirements Coverage (79 total)

All 79 requirements (R01-R79) have code written and deployed as Docker images. NONE are integration-tested. See deployment-snapshot-20260208.md for full mapping.

When to Use This Skill

1. System Architecture: Understanding service topology, communication patterns
2. Service Dependencies: Which services interact and how
3. Technology Decisions: Tech stack, databases, messaging
4. Integration Planning: Designing APIs or workflows
5. Deployment Status: What's running where
6. Stakeholder Communication: High-level system explanations

Related Skills

• flowmaster-backend — API endpoints, service contracts
• flowmaster-server — Infrastructure, deployment, CI/CD
• flowmaster-frontend — UI components, patterns
• flowmaster-database — ArangoDB schema, collections
• flowmaster-environment — Service env vars, ports, config
• flowmaster-tools — MCP tools, integrations, SDX