Deployment & Infrastructure

MFC Safe Pty Ltd PaaS Deployment & Infrastructure Operations Manual Title: Deployment and Hosting Guide for the MFC Safe Platform Version: 1.0 Effective Date: 2 June 2025 Next Review Date: 2 June 2026

---

1. Purpose

This manual outlines the deployment, hosting, and operational processes for the MFC Safe Platform-as-a-Service (PaaS) infrastructure. It defines the continuous integration and delivery (CI/CD) workflow, hosting architecture on AWS, scaling strategies, service provisioning, and related DevOps procedures.

---

2. Cloud Hosting Environment

• Primary Provider: Amazon Web Services (AWS)

• Regions Used: ap-southeast-2 (Sydney), ap-southeast-1 (Singapore)

• Availability Zones: Multi-AZ deployments for redundancy

• Core Services:

  • EC2 (compute)

  • RDS / MongoDB Atlas (databases)

  • S3 (object storage)

  • CloudWatch (monitoring & logs)

  • IAM (role-based access)

  • Secrets Manager (credential storage)

  • API Gateway / Lambda (serverless APIs)

---

3. Infrastructure Provisioning

• All infrastructure is defined using Terraform (Infrastructure-as-Code)

• Code repositories are version controlled via GitHub

• Environments: dev, staging, prod

• Terraform modules provision:

  • VPCs and subnets

  • Security groups and firewalls

  • EC2 launch templates

  • CloudWatch alarms and SNS topics

---

4. CI/CD Pipeline

• CI Platform: GitHub Actions

• Stages:

  • Linting & Unit Tests

  • Security Scan (Snyk / CodeQL)

  • Build Docker images

  • Push to Amazon ECR

  • Deploy via Helm to Kubernetes (EKS)

• Release Strategy:

  • Feature branches → staging → main (production)

  • Canary releases supported via label-based service selectors

  • Manual approval required for production deployment

---

5. Monitoring & Logging

• Application logs: CloudWatch (ECS), Sentry (frontend)

• Infrastructure logs: AWS CloudTrail

• Alerts: SNS + Slack integration

• Dashboards: Grafana for metrics (CPU, memory, I/O, response times)

---

6. Scaling & Resilience

• Horizontal scaling via Kubernetes Horizontal Pod Autoscaler (HPA)

• LoRaWAN gateways and sensor ingestion services are decoupled via queueing (SQS)

• MongoDB replica sets for HA, with daily and weekly backups

• Load balancers with health check routing (ALB)

---

7. Service Provisioning

• Onboarding a new client site includes:

  • Namespace creation (Kubernetes)

  • Subdomain registration and SSL provisioning

  • Database schema initialization

  • API credentials for data ingestion

  • Optional: custom webhook configurations

---

8. Backup & Recovery

• Automated daily and weekly backups

• Retention: 90 days

• Backup verification tests monthly

• RTO target: < 4 hours; RPO target: < 1 hour

---

Maintained By: DevOps & Infrastructure Team Contact: [e-mail] Approved By: Technology Director, MFC Safe Pty Ltd

---

End of Document