How CloudJournee migrated a social shopping platform from a single VPS to a scalable, multi-tier AWS architecture in 5 weeks
About the Customer
Our client is a fast-growing digital commerce platform operating in the social marketplace ecosystem, enabling D2C brands and emerging sellers to drive engagement through video-led shopping experiences. The platform combines e-commerce with short-form visual content, allowing homegrown Indian brands and D2C sellers to showcase products through video-centric listings while shoppers discover, follow, and purchase from their favourite brands through personalised feeds and seamless checkout experiences.
The platform was experiencing rapid growth, handling increasing user traffic and engagement across mobile-first audiences in India. However, the entire application stack — React frontend, Node.js backend, and MongoDB database — was hosted on a single Hostinger VPS, with all components competing for the same compute, memory, disk, and network resources.
Business Challenge
The platform’s legacy infrastructure, built entirely on a single VPS, created significant operational and business risks that directly impacted growth, user experience, and revenue.
| Challenge | Business Impact | Quantified Loss |
|---|---|---|
| Frequent downtime | Any single failure brought the entire system offline, affecting user trust and active transactions | 3-10 hours/month of total platform outage |
| Performance bottlenecks | Frontend, backend, and database sharing the same resources caused degraded response times under normal load | 2.8s page loads, 1.2s API response times |
| Revenue loss during peak traffic | Marketing campaigns drove traffic spikes that the single server could not absorb, resulting in crashed sessions and abandoned carts | ~40% of paid campaign traffic lost during peak periods |
| Critical data risk | MongoDB hosted on the same disk as the application with no backups, no replication, and no disaster recovery | High risk of permanent, unrecoverable data loss |
| No security controls | No encryption at rest or in transit, no network segmentation, no WAF, no access controls beyond SSH | Platform exposed to breaches and compliance gaps |
| Scalability ceiling | Limited to vertical scaling on a single server with no path to horizontal scaling or multi-AZ resilience | Could not onboard new brand partnerships without infrastructure risk |
CloudJournee Assessment & Approach
CloudJournee conducted a structured assessment of the existing platform covering application architecture, workload patterns, data dependencies, and security posture. The assessment identified that a lift-and-shift migration would not resolve the fundamental issues — the single-server monolithic architecture had to be decomposed into a multi-tier, cloud-native design to address the scalability, resilience, and security gaps.
CloudJournee designed and executed a phased migration strategy:
| Phase | Activities | Deliverables | Duration |
|---|---|---|---|
| 1. Assess | Architecture review, workload profiling, dependency mapping, security gap analysis | Migration plan, target architecture design, risk assessment | Week 1 |
| 2. Foundation | AWS landing zone setup, multi-account structure, VPC with public/private/database subnets across 2 AZs, IAM roles, encryption | Secure cloud foundation with network segmentation and access controls | Week 2 |
| 3. Migrate | Frontend to AWS Amplify + CloudFront, backend to EC2 Auto Scaling Group behind ALB, MongoDB to DocumentDB Multi-AZ, DNS cutover via Route 53 | Production-ready multi-tier architecture with data validation and rollback capability | Weeks 3-4 |
| 4. Optimise | Performance tuning, CloudWatch monitoring and alerting, WAF rules, load testing, documentation and handover | Operational runbook, monitoring dashboards, go-live sign-off | Week 5 |
Solution: Multi-Tier Cloud-Native Architecture
CloudJournee redesigned the platform into a three-tier, cloud-native architecture deployed in the AWS Mumbai region (ap-south-1), with complete separation of concerns enabling each layer to scale independently.
Architecture Overview
| Layer | Before (Single VPS) | After (AWS Multi-Tier) |
|---|---|---|
| Frontend | React app served from same VPS, no CDN, no caching | AWS Amplify with CloudFront CDN (13 edge locations in India), HTTPS via ACM |
| Backend / API | Node.js on same VPS, no load balancing, no auto-scaling, single process | EC2 Auto Scaling Group behind Application Load Balancer, multi-AZ, scales 2-10 instances based on CPU/request metrics |
| Database | MongoDB on same disk as app, no replication, no backups, no failover | Amazon DocumentDB (MongoDB-compatible) Multi-AZ with automated daily backups (35-day retention), point-in-time recovery, encryption at rest via KMS |
| Networking | Single public IP, flat network, no segmentation | VPC with public, private, and database subnets across 2 AZs, NAT Gateway, Route 53 DNS with health checks and failover routing |
| Security | No encryption, no WAF, no IAM, SSH-only access | AWS WAF + Shield, IAM Identity Center, KMS encryption (at rest and in transit), Secrets Manager, Certificate Manager, CloudTrail audit logging |
| Monitoring | None — issues discovered by users reporting errors | CloudWatch metrics, alarms, and dashboards, CloudTrail for audit, Systems Manager for patching, proactive alerting via SNS |
AWS Services Deployed
| Category | Services |
|---|---|
| Frontend & Compute | AWS Amplify, Amazon EC2, Auto Scaling Group |
| Database | Amazon DocumentDB (MongoDB-compatible), Multi-AZ |
| Networking & CDN | Amazon VPC (multi-AZ), Application Load Balancer, Amazon CloudFront, Amazon Route 53 |
| Security | AWS WAF & Shield, AWS IAM & Identity Center, AWS KMS, Secrets Manager, AWS Certificate Manager |
| Operations & Storage | Amazon CloudWatch, AWS CloudTrail, AWS Systems Manager, Amazon S3 |
Results
The migration delivered measurable improvements across every dimension of platform performance, reliability, and business outcomes:
| Metric | Before | After | Improvement |
|---|---|---|---|
| Platform availability | ~96% (~3-10 hrs downtime/month) | 99.9% | >99.9% uptime |
| Page load time | 2.8 seconds | 1.1 seconds | 50% faster |
| API response time | 1.2 seconds | 350 milliseconds | 70% faster |
| Checkout conversion rate | ~65% | ~82% | 25% increase |
| User retention | ~55% | ~68% | 20% improvement |
| Campaign traffic handling | 60% (40% lost to crashes) | 100% handled | Full marketing ROI captured |
| Engineering firefighting | 30% of engineering time | 5% | 50% reduction in ops overhead |
| Scalability | Single server, vertical only | Auto Scaling Group, multi-AZ | 3-5x growth capacity |
| Data protection | No backups, no DR | Automated daily backups, multi-AZ failover, PITR | Fully secured |
| Security posture | None | Enterprise-grade (WAF, KMS, IAM, CloudTrail) | 70% risk reduction |
Customer Testimonial
“The migration completely transformed our platform performance and reliability. What used to be a constant struggle with downtime and slow response is now a stable, high-performing system that supports our growth seamlessly. The impact on user experience and conversions has been immediate and measurable.”
— CTO, Mobile Marketplace Platform
Business Value Delivered
Operational Excellence
CloudJournee eliminated platform downtime and established enterprise-grade reliability. With auto-scaling and load balancing, the platform handles traffic spikes effortlessly, ensuring consistent performance and seamless user experiences during marketing campaigns and organic growth periods.
Revenue Growth
Improved speed and availability translated directly into higher conversions and better campaign performance. With zero traffic loss during peak loads, the client’s marketing investments now deliver full ROI — an estimated 40% recovery in previously lost campaign revenue.
Cost Optimisation
By replacing an over-provisioned single VPS with right-sized, auto-scaling infrastructure, the platform reduced operational overhead while maximising resource utilisation. Engineering effort shifted from 30% firefighting to 5%, freeing the team to focus on product development.
Security & Compliance
Enterprise-grade security — including encryption at rest and in transit (KMS), web application firewall (WAF), identity and access management (IAM Identity Center), and audit logging (CloudTrail) — ensured complete protection of user data and transaction integrity.
Scalability & Future Readiness
The platform is now built to scale dynamically from its current user base to 3-5x growth without infrastructure changes. This positions the business for rapid expansion, new brand partnerships, and geographic growth across India.
About CloudJournee
CloudJournee is an AWS Advanced Tier Partner based in Bengaluru, India. We specialise in cloud migration, modernisation, and AI/ML delivery for startups, mid-market companies, and enterprises across India and globally.
Our team brings deep expertise in AWS infrastructure, containerisation, serverless architectures, and data platform modernisation — delivering outcomes like this case study: measurable business impact, on time, with zero downtime during migration.
Ready to modernise your platform?
CloudJournee helps startups and mid-market companies migrate to AWS with enterprise-grade architecture — without enterprise complexity.
