Reactive Shipping System

A reactive, event-driven warehouse fulfilment platform. Orders are ingested, routed to the optimal fulfilment centre, picked, packed, labelled, and handed off to carriers — all driven by Kafka events and asynchronous microservices.

Documentation

Document	Description
PRD	Product requirements and business context
SDS	System design — architecture, data models, event contracts
Implementation Plan	48-week phased delivery plan across 10 squads
Initial Thoughts	Early design notes and trade-offs

High-level architecture

Customer / OMS
      │
      ▼
 Kong Gateway
      │
      ▼
Order Ingestion ──► FC Router ──► Inventory Service
                                        │
                                   Wave Planner
                                        │
                                   Pick Engine
                                        │
                                  Pack Service ◄── (this repo, Sprint 5)
                                        │
                              Carrier & Tracking
                                        │
                                  [OrderShipped]

Services

Service	Location	Language	Status
Pack Service	services/pack-service	Clojure	🟡 In progress

More services will be added here as each squad scaffolds their domain (see Implementation Plan).

Tech stack

Concern	Technology
Messaging	Apache Kafka 3.8 (KRaft) + Confluent Schema Registry
Primary DB	ScyllaDB 6 (wide-row, low-latency)
Cache	Valkey 8 (Redis-compatible)
Search	OpenSearch 2.x
Analytics	Apache Trino + MinIO
Service mesh	Istio 1.23 (mTLS)
API gateway	Kong Gateway OSS
Observability	Prometheus 3 + Grafana 11 + Jaeger 2 + OpenTelemetry
CI/CD	GitHub Actions + ArgoCD + Harbor
Secrets	HashiCorp Vault
Feature flags	Unleash