Infrastructure Environments
Deployment Environments
Webforx Platform Engineering operates three isolated environments. Code flows through Stage, Sandbox, and Production in a controlled promotion pipeline.
Stage
promote
Sandbox
promote
Production
Development & Integration
Stage
The first deployment target for new code. Used by developers for integration testing, feature validation, and early bug detection. Receives every merged PR automatically.
Purpose: Integration testing, feature validation, API contract verification
Deploy trigger: Automatic on merge to develop branch
Data: Synthetic/seeded test data, reset periodically
Stability: Lowest -- frequent deploys, may break
Access: Development team, QA engineers
Services: All 12 microservices + Kong gateway + Redis + RabbitMQ + PostgreSQL
Pre-Production & QA
Sandbox
Mirror of production used for final QA, UAT, performance testing, and stakeholder demos. Promoted code from Stage lands here for thorough validation before going live.
Purpose: UAT, performance testing, stakeholder demos, release candidate validation
Deploy trigger: Manual promotion from Stage after QA sign-off
Data: Production-like dataset, anonymized where needed
Stability: Medium -- should mirror prod behavior closely
Access: QA team, product owners, selected stakeholders
Infra: Production-equivalent sizing, same Kubernetes cluster config
Live & Customer-Facing
Production
The live environment serving real users and customers. Only thoroughly tested and approved releases from Sandbox are promoted here. Zero-downtime deployments enforced.
Purpose: Live customer traffic, real transactions, production workloads
Deploy trigger: Manual promotion from Sandbox with release approval
Data: Real customer data, encrypted at rest, backups on schedule
Stability: Highest -- zero-downtime deploys, rollback ready
Access: Platform engineering (deploy), on-call team (ops), limited SSH
Infra: HA Kubernetes, multi-AZ, auto-scaling, Vault secrets, monitoring + alerting
Infrastructure Comparison
| Component | Stage | Sandbox | Production |
|---|---|---|---|
| Kubernetes Nodes | 2 nodes (t3.medium) | 3 nodes (t3.large) | 3+ nodes (t3.xlarge), auto-scale |
| PostgreSQL | Single instance (db.t3.micro) | Single instance (db.t3.small) | Multi-AZ RDS (db.t3.medium+) |
| Redis | Single node | Single node with replica | Redis Sentinel (master + 2 replicas + 3 sentinels) |
| RabbitMQ | Single node | Single node | Clustered (3 nodes), mirrored queues |
| S3 / MinIO | MinIO (local) | AWS S3 | AWS S3 (versioned, lifecycle rules) |
| Kong Gateway | Single instance | Single instance | 2+ replicas, health checks |
| Monitoring | Basic logs | Prometheus + Grafana | Prometheus + Grafana + PagerDuty alerting |
| Secrets | Env files / .env | Vault (dev policy) | Vault (production policy, audit logging) |
| Backups | None | Daily snapshots | Continuous + point-in-time recovery |
| CORS Origins | localhost:3000, localhost:3010 | stage.*, sandbox.* domains | Production domains only |
| TLS | Self-signed / none | Let's Encrypt | ACM managed certificates |
| Deploy Frequency | Multiple per day | Weekly / on-demand | Bi-weekly / release cycle |