CASE STUDY

Novagems: SaaS infrastructure that runs 10,000+ field workers.

Multi-tenant workforce management SaaS serving security and cleaning companies across the United States and Canada — built on AWS EKS, Kubernetes, and Terraform.

10,000+

Active field workers

2

Countries (US + CA)

99.9%

Uptime SLA

3 yrs

Ongoing retainer

THE CHALLENGE

One platform. Many tenants. Zero downtime tolerance.

Novagems started as a security guard scheduling tool. Within two years, clients needed real-time GPS tracking, payroll integration, incident reporting, and multi-company tenancy — all while the platform stayed online 24/7 for field supervisors across time zones.

A monolithic architecture and a single-tenant data model couldn’t absorb this growth. The team needed to re-platform without taking the product offline.

Single-tenant DB

Each client had their own database. Adding a client meant a manual provisioning process. Didn’t scale past 20 accounts.

No CI/CD pipeline

Deployments were manual SSH sessions. Rollbacks took hours. Dev and production environments drifted constantly.

Infrastructure drift

EC2 instances were manually configured. No infrastructure-as-code meant rebuilding from memory after any failure.

THE SOLUTION

Multi-tenant on EKS. Infrastructure as code from day one.

We re-architected Novagems around a shared-schema multi-tenant model with row-level tenant isolation, then containerised every service and moved the entire stack to AWS EKS managed by Terraform. GitOps with ArgoCD replaced manual deployments.

AWS EKSKubernetesTerraformArgoCDGitHub ActionsPostgreSQLRedisNode.jsReact NativeCloudflare

WHAT WE DELIVERED

Eight capabilities shipped across 18 months.

Multi-tenant architecture

Shared PostgreSQL schema with row-level tenant isolation. Adding a new client: 30 seconds, zero manual steps.

EKS cluster

Kubernetes cluster across 3 AZs. Auto-scaling node groups. Zero-downtime deployments via rolling updates.

Terraform IaC

All infrastructure defined in code. Reproducible from scratch in under 20 minutes. Drift detection in CI.

GitOps pipeline

GitHub Actions → ArgoCD. Every PR gets a preview environment. Production deploys require PR approval.

Real-time GPS tracking

WebSocket-based live location feed for field supervisors. Works on 3G. Battery-optimised native app.

Payroll & scheduling

Shift scheduling engine with conflict detection. Direct payroll export to ADP and Paylocity formats.

Incident reporting

Mobile-first incident capture with photo upload, GPS tagging, and automatic escalation routing.

Monitoring stack

CloudWatch + PagerDuty. On-call runbooks. 99.9% uptime SLA maintained for 3 consecutive years.

Build something like this.

Building a SaaS product that needs to scale without breaking? We’ll audit your current stack and give you a fixed quote within 3 working days.


Tell us about your project