skip to content
Blake Green

Thank You Serverless!

/ 4 min read

aws, serverless, infrastructure, saas

Thank You Serverless: How I Scaled RapidEPR to 250k Users for $100/Month

As a SaaS founder, one of the most critical decisions you’ll make is choosing your infrastructure approach. Today, I want to share how AWS serverless technology has been a game-changer for RapidEPR, allowing us to scale to over 250,000 users while keeping our monthly infrastructure costs around $100 and virtually eliminating operational overhead.

The Serverless Promise vs. Reality

When AWS first introduced Lambda in 2014, the serverless promise was enticing: no servers to manage, pay only for what you use, and automatic scaling. But many developers remained skeptical. Would it really work for production applications? Could it handle real-world scale? Would costs remain predictable?

After several years of running RapidEPR on a serverless architecture, I can confidently say: Yes, serverless delivers on its promises. And it does so in ways that have fundamentally transformed how I think about building and scaling SaaS applications.

RapidEPR’s Serverless Architecture

Before diving into the benefits, let me outline our architecture:

Frontend

  • Next.js application hosted on AWS S3
  • CloudFront for global content delivery and caching
  • Static assets and dynamic content served with minimal latency worldwide

Authentication

  • Amazon Cognito handling all user authentication and authorization
  • Secure user pools and identity management without custom auth code
  • Social login integration and MFA support

API Layer

  • API Gateway serving as the entry point for all client requests
  • RESTful endpoints mapped to specific Lambda functions
  • Request validation, throttling, and API keys managed automatically

Business Logic

  • Lambda functions written in Node.js handling all application logic
  • Separate functions for different domains (user management, data processing, reporting)
  • Event-driven architecture for asynchronous processing

Data Storage

  • DynamoDB as our primary NoSQL database
  • Single-digit millisecond response times even at scale
  • On-demand capacity mode adjusting automatically to traffic patterns

The $100/Month Miracle

When I tell other founders that our infrastructure costs hover around $100/month despite serving 250,000 users, they’re often in disbelief. Here’s how the costs break down:

  • Lambda: ~$20/month (millions of invocations, but most complete in less than 100ms)
  • API Gateway: ~$15/month
  • DynamoDB: ~$40/month (on-demand pricing)
  • CloudFront/S3: ~$15/month
  • Cognito: ~$10/month

The beauty of serverless is that costs scale linearly with usage. We’re not paying for idle servers or over-provisioned resources. During quiet periods, our costs drop even lower.

Zero Operational Overhead

Perhaps even more valuable than the cost savings is the operational freedom. With a traditional architecture, I’d need to:

  • Monitor server health
  • Apply security patches
  • Handle load balancing
  • Manage database scaling
  • Worry about high availability
  • Plan for disaster recovery

With our serverless setup, AWS handles all of this. I haven’t SSH’d into a server in years. I don’t wake up to alerts about CPU spikes or memory leaks. I don’t stress about scaling for traffic surges.

This operational simplicity has allowed our small team to focus entirely on product development and customer success, rather than infrastructure management.

Scaling Challenges We Overcame

The serverless journey wasn’t without challenges. Here are a few we encountered and solved:

Cold Starts

Initially, Lambda cold starts caused occasional latency spikes. We mitigated this by:

  • Optimizing function size and dependencies
  • Using Provisioned Concurrency for critical paths
  • Implementing client-side retry strategies

DynamoDB Design

NoSQL database design requires a different mindset. We learned to:

  • Design access patterns first, then create tables and indexes
  • Use composite keys effectively
  • Implement sparse indexes for filtering

Monitoring and Debugging

Distributed systems can be harder to debug. We addressed this by:

  • Implementing comprehensive logging with AWS CloudWatch
  • Using X-Ray for tracing requests across services
  • Creating custom dashboards for key metrics

Would I Choose Serverless Again?

Absolutely, without hesitation. The combination of cost efficiency, operational simplicity, and automatic scaling has been transformative for our business. We’ve been able to grow from 10 users to 250,000 without infrastructure becoming a bottleneck or a financial drain.

For startups and growing SaaS businesses, serverless isn’t just a technical choice—it’s a strategic advantage. It allows you to focus resources on what truly matters: building a product your customers love.

Final Thoughts

If you’re building a new SaaS application or considering a migration for an existing one, I strongly encourage you to explore serverless architectures. The learning curve is real, but the long-term benefits far outweigh the initial investment.

AWS serverless technology has been a faithful partner in RapidEPR’s growth journey, and for that, my team, my customers, and especially my wallet say: Thank you, Serverless.

Found this insightful? If you're interested in my AWS consulting, please reach out to me via email or on X