How We Scaled a SaaS Platform to 1 Million Requests/Day

Client: ProjectHub

1M+

Daily Requests

99.99%

Uptime

60%

Query Speed Improvement

$50K

Annual Savings

10K+

Concurrent Users

Zero

Deployment Downtime

The Challenge

ProjectHub's project management SaaS was hitting performance limits at 100K daily requests. The monolithic Node.js backend on a single server experienced frequent downtime during peak hours, slow database queries, and a growing queue of user complaints about performance.

Our Solution

We re-architected the platform with edge computing, database optimization, and intelligent caching. The backend was split into microservices running on Cloudflare Workers, the database was optimized with proper indexing and read replicas, and a multi-layer caching strategy was implemented.

Implementation Details

Decomposed the monolithic backend into 12 microservices on Cloudflare Workers

Optimized PostgreSQL with proper indexing, query optimization, and connection pooling

Implemented Redis caching layer for session data and frequently accessed resources

Built a CDN strategy for static assets with Cloudflare R2

Added WebSocket support via Durable Objects for real-time collaboration

Implemented database read replicas for read-heavy endpoints

Set up auto-scaling with load-based routing across edge locations

Built comprehensive monitoring with custom Cloudflare analytics

Results

10x traffic capacity (100K to 1M+ daily requests)

99.99% uptime (was 97%)

60% faster database queries

$50K/year infrastructure savings

Real-time collaboration for 10K+ concurrent users

Zero downtime deployments

Technology Stack

Next.js

Cloudflare Workers

PostgreSQL

Redis

Durable Objects

TypeScript

Docker

“Sundarban Lab took our struggling platform and turned it into a scalable, reliable system. We went from constant downtime to 99.99% uptime, and our users immediately noticed the performance improvement.”

Sarah Johnson

CEO, ProjectHub