Skip to content

Latest commit

 

History

History
39 lines (30 loc) · 1.23 KB

File metadata and controls

39 lines (30 loc) · 1.23 KB

Cost Model & Optimization

This project is built around wake-on-demand architecture to minimize AWS cost.

Cost While Awake

Approximate monthly cost if the environment stays online continuously:

  • EC2 ASG: $10–15/month
  • RDS t3.micro: $12–15/month
  • S3 + CloudFront: < $1/month
  • Lambda + API GW: negligible (< $0.10)
  • SSM Parameters: free (standard tier)

Total always-on ≈ $25–32/month

Cost While Sleeping

With autosleep after X minutes:

  • EC2 → $0
  • RDS → $0 (destroyed)
  • Only S3/CF/SSM remain → < $1/month

Wake Overhead

Each wake event incurs the following operational cost and time:

  • GitHub Actions runner time: ~1–2 minutes
  • Terraform operations: 3–5 minutes
  • AWS service cold starts:
    • EC2 ASG instance warm-up: 60–120 seconds
    • RDS initialization: 2–4 minutes
  • Total wake time: typically 8–15 minutes
    (from pressing “Wake” to the application responding over HTTPS)

There is no persistent cost once the idle reaper destroys all resources.

Key Optimization Principles

  • Destroy infrastructure on inactivity
  • Stateless frontend via S3
  • OIDC (no permanent IAM users)
  • Auto-scaling app layer only during active sessions