Services Overview

This section provides detailed documentation for each infrastructure service in the HealthFlow NDP platform.

Infrastructure Services

Gateway Layer

Traefik

Purpose: Ingress controller and API gateway
Stack: Gateway
Port: 80, 443, 8080 (dashboard)

→ Full Documentation

Key features:

• Automatic HTTPS with Let's Encrypt
• Dynamic routing and load balancing
• Middleware chain (auth, rate-limiting, CORS)
• Prometheus metrics export

---

Portainer

Purpose: Kubernetes management UI
Stack: Gateway
Port: 9000

→ Full Documentation (Coming soon)

Key features:

• Visual container management
• RBAC and team management
• Stack deployment interface
• Log viewer and console access

---

Data Layer

PostgreSQL

Purpose: Primary relational database
Stack: Data
Port: 5432

→ Full Documentation

Usage:

• Prescription data
• Patient registry
• HPR (Healthcare Provider Registry)
• Dispense records

---

MySQL

Purpose: Legacy database for specific registries
Stack: Data
Port: 3306

→ Full Documentation

Usage:

• Pharmacy registry
• Medicine directory
• Legacy EDA integration

---

Redis

Purpose: In-memory cache and session store
Stack: Data
Port: 6379

→ Full Documentation

Usage:

• API response caching
• Session management
• Rate limiting counters
• Real-time data

---

MongoDB

Purpose: Document database for audit logs
Stack: Data
Port: 27017

→ Full Documentation (Coming soon)

Usage:

• Audit trail storage
• Compliance logging
• Analytics data

---

Monitoring Layer

Prometheus

Purpose: Metrics collection and storage
Stack: Monitoring
Port: 9090

→ Full Documentation

Metrics sources:

• Kubernetes cluster metrics
• Application metrics
• Infrastructure metrics
• Custom business metrics

---

Grafana

Purpose: Metrics visualization and dashboards
Stack: Monitoring
Port: 3000

→ Full Documentation

Dashboards:

• Cluster health
• Service performance
• Business KPIs
• Alert overview

---

Loki

Purpose: Log aggregation system
Stack: Monitoring
Port: 3100

→ Full Documentation

Features:

• Centralized log storage
• Label-based indexing
• Grafana integration
• Cost-effective storage

---

AlertManager

Purpose: Alert routing and notifications
Stack: Monitoring
Port: 9093

→ Full Documentation (Coming soon)

Channels:

• Email notifications
• Slack webhooks
• PagerDuty integration
• SMS alerts

---

Service Discovery Layer

Consul

Purpose: Service mesh and service discovery
Stack: Discovery
Port: 8500 (UI), 8600 (DNS)

→ Full Documentation

Features:

• Service registration
• Health checking
• KV store
• Service mesh (sidecar proxies)

---

Vault

Purpose: Secrets management
Stack: Discovery
Port: 8200

→ Full Documentation

Secrets:

• Database credentials
• API keys
• TLS certificates
• Encryption keys

---

Service Communication Matrix

Service Dependencies

Service	Depends On	Required For
Traefik	None	All external access
Consul	None	Service discovery, Vault
Vault	Consul	All applications
PostgreSQL	None	Core applications
Redis	None	Caching, sessions
Prometheus	None	Monitoring
Grafana	Prometheus, Loki	Dashboards
Loki	None	Log aggregation
Applications	PostgreSQL, Redis, Consul, Vault	End-user features

Port Allocation

Gateway Stack

Service	Port	Protocol	Exposure
Traefik	80	HTTP	External
Traefik	443	HTTPS	External
Traefik	8080	HTTP	Internal
Traefik	8082	HTTP	Internal (metrics)
Portainer	9000	HTTP	Internal

Data Stack

Service	Port	Protocol	Exposure
PostgreSQL	5432	TCP	Internal
MySQL	3306	TCP	Internal
Redis	6379	TCP	Internal
MongoDB	27017	TCP	Internal
Kafka	9092	TCP	Internal

Monitoring Stack

Service	Port	Protocol	Exposure
Prometheus	9090	HTTP	Internal
Grafana	3000	HTTP	Internal
Loki	3100	HTTP	Internal
AlertManager	9093	HTTP	Internal
Node Exporter	9100	HTTP	Internal

Discovery Stack

Service	Port	Protocol	Exposure
Consul	8500	HTTP	Internal
Consul	8600	DNS	Internal
Vault	8200	HTTPS	Internal

Resource Requirements by Service

Important Note

All resource specifications below are rough estimates for initial deployment. Production requirements should be determined based on:

• Load testing results
• Actual traffic patterns
• Data volume and growth rate
• Performance monitoring and profiling

Start with these baseline values and use Horizontal Pod Autoscaling (HPA) and Vertical Pod Autoscaling (VPA) for automatic adjustments.

High Priority (Critical)

Service	CPU	Memory	Storage	Notes
PostgreSQL	4 cores	8 GB	500 GB	Primary database - scale based on query load
Redis	2 cores	4 GB	20 GB	Cache layer - increase for higher throughput
Traefik	2 cores	2 GB	10 GB	Gateway - scale with request volume
Consul	2 cores	2 GB	20 GB	Service mesh - depends on service count

Medium Priority (Important)

Service	CPU	Memory	Storage	Notes
Prometheus	2 cores	4 GB	100 GB	Metrics storage - adjust retention period
Vault	1 core	1 GB	10 GB	Secrets - minimal resource needs
MySQL	2 cores	4 GB	100 GB	Legacy data - estimate based on dataset
Grafana	1 core	2 GB	10 GB	Dashboards - light workload

Low Priority (Supporting)

Service	CPU	Memory	Storage	Notes
Loki	1 core	2 GB	200 GB	Logs - storage depends on retention
AlertManager	0.5 core	512 MB	5 GB	Alerts - minimal resources
Portainer	0.5 core	512 MB	10 GB	Management - UI only

Service Health Checks

All services implement three types of health checks:

1. Liveness Probe

Determines if the container is running. Failed checks result in container restart.

2. Readiness Probe

Determines if the service can accept traffic. Failed checks remove the pod from service endpoints.

3. Startup Probe

Used for slow-starting applications. Disables liveness/readiness checks until first success.

Example:

livenessProbe:
  httpGet:
    path: /health/live
    port: 8080
  initialDelaySeconds: 30
  periodSeconds: 10
  timeoutSeconds: 5
  failureThreshold: 3

readinessProbe:
  httpGet:
    path: /health/ready
    port: 8080
  initialDelaySeconds: 10
  periodSeconds: 5
  timeoutSeconds: 3
  failureThreshold: 3

Next Steps

Explore detailed documentation for each service:

• Traefik Gateway
• PostgreSQL
• Redis
• Prometheus
• Grafana
• Loki
• Consul
• Vault

Or continue with stack deployment:

• Gateway Stack
• Data Stack
• Monitoring Stack