Skip to content
SaaS2024-07

Pre-Seed Startup: Production Infrastructure in 2 Weeks, $400/Month

A two-person pre-seed startup had a working product but was running it on a personal laptop during demos. They needed a production deployment before a fundraising pitch in 3 weeks. We built a lean, investor-ready infrastructure stack for $400/month.

Deploy Time
Manual - founder SSHes into server
4 minutes CI/CD on merge
Deploy Frequency
Once a week (scary)
Multiple times daily
Incidents
No monitoring, no alerts, no uptime visibility
Full uptime monitoring, alerting, and deployment visibility
Cost Impact
-
$387/month total infrastructure (vs $2K+ typical early-stage stacks)

The Challenge

The constraint was cost - they had $60K left in the bank and 6 months to close a seed round. Every dollar spent on infrastructure was a dollar not spent on product. But they also needed to demonstrate production reliability in the pitch: uptime, monitoring, security basics, and a deployment process that did not require a founder to SSH into a server.

The Approach

We built for the current scale - 200 users, <1,000 requests/day - not for hypothetical future scale. One ECS Fargate service, one RDS instance (db.t4g.small), one S3 bucket, GitHub Actions CI/CD, and a basic Grafana Cloud monitoring setup. Infrastructure-as-code from day one so it can grow without rewriting.

The Implementation

Lean AWS stack via Terraform

VPC with two availability zones, ECS Fargate (0.25 vCPU / 0.5GB RAM), RDS PostgreSQL db.t4g.small with a 7-day automated backup, ACM certificate, ALB, and Route 53 for the custom domain. Total monthly cost at launch: $387/month. Founders can deploy by merging a pull request.

AWS ECS FargateAWS RDSTerraformRoute 53

GitHub Actions CI/CD pipeline

On merge to main: run tests, build Docker image, push to ECR, deploy to ECS. The full pipeline takes 4 minutes. No human intervention required. Founders ship features by merging PRs - no DevOps knowledge required.

GitHub ActionsAWS ECRAWS ECS

Secrets and configuration management

All secrets in AWS Secrets Manager, injected at container startup via ECS task role. No environment variables in the codebase. The application startup fails fast if a required secret is missing - no silent misconfiguration.

AWS Secrets ManagerAWS IAMECS Task Roles

Monitoring and uptime alerting

Grafana Cloud free tier for basic application metrics. AWS CloudWatch for ECS and RDS. UptimeRobot for external uptime monitoring with email + Slack alerts. Three dashboards: overall health, database performance, and deployment history. Total monitoring cost: $0/month.

Grafana CloudAWS CloudWatchUptimeRobot

Key Takeaways

  • Pre-seed infrastructure should be optimised for cost and simplicity, not scalability - right-sizing early saves the runway that matters
  • Terraform from day one costs 4 hours but pays back every time you need to change something - infrastructure-as-code is not just for big teams
  • ECS Fargate + RDS is the right stack for early-stage SaaS - no EC2 management, pay only for what you use
  • Free monitoring (Grafana Cloud free tier + UptimeRobot) is entirely sufficient for 0–10K users - premium monitoring tools are a premature optimisation

Facing Similar Challenges?

Book a free 30-minute audit and I will tell you what I see.

Book Free Audit
All case studies