Draw Your Architecture, Get Your AWS Monthly Cost — I Built and Published a Tool for That
I built a free web tool that visually estimates AWS serverless costs by connecting service nodes in a flow editor — automatically calculating implicit costs like CloudWatch Logs and propagating request counts downstream.
Note: This article was originally written in Japanese and translated. Some expressions may sound unnatural — feedback is welcome!
Introduction
Designing AWS architectures with serverless resources is genuinely fun.
However, the most tedious part is calculating the monthly cost of that architecture. I work as a cloud solutions architect at a systems integrator, and whenever I design an architecture, a cost estimate is always expected alongside it. Often, I need to produce estimates when access patterns and data volumes are still uncertain, and it's not uncommon to be asked for estimates across multiple traffic scenarios.
The most popular estimation tool is the official AWS Pricing Calculator .
However, this tool requires you to add services one by one, which makes it time-consuming to get a sense of the overall cost. Additionally, in a standard Lambda configuration, logs are sent to CloudWatch Logs, incurring implicit costs — and factoring these into your estimate is extra work.
So I built Flowdget — a tool that lets you visually assemble your architecture in a flow editor while estimating costs in real time!
What is Flowdget?
Flowdget is a web application that visually estimates the cost of AWS serverless architectures using a flow editor.
Flow + Budget = Flowdget.
Features
- Flow editor — Drag and drop nodes, connect them with arrows to build your architecture
- Automatic request propagation — Upstream request counts automatically flow downstream (multiplier configurable)
- Implicit cost calculation — Adding Lambda automatically shows CloudWatch Logs costs
- Cache hit rate awareness — CloudFront cache hit rate adjusts downstream request counts
- Free Tier ON/OFF — Compare production (no free tier) vs development (with free tier) costs
- Region switching — Instantly reflect pricing differences between Tokyo / N. Virginia
- Save & restore — JSON export/import, auto-save to browser storage
Deep Dive
1. Implicit Cost Visualization
When you add a Lambda node, CloudWatch Logs ingestion + storage costs are automatically calculated and displayed. For WAF and CloudFront, you can toggle log destinations (CloudWatch Logs / S3) with checkboxes.
2. Automatic Request Propagation
When you change the request count on an upstream node, all downstream nodes update automatically. Set edge multipliers (×0.5, ×2, etc.) to express fan-out architectures.
Set CloudFront's cache hit rate to 80%, and only 20% of requests propagate downstream.
3. Automatic Pricing Data Updates
A Lambda batch job fetches the latest pricing weekly from the AWS Price List API. Pricing changes are automatically reflected.
4. Edge Thickness Visualizes Traffic Volume
Edge (connection line) thickness scales proportionally to request count, making it easy to see which paths carry the most traffic at a glance.
5. Export Function
You can export your configuration. Node positions, connections, and parameters are fully saved and can be restored via import.
Sample Architectures
EventBridge + Step Functions + Bedrock API
Differences in execution counts from Step Functions branching are automatically tracked.
SPA (CloudFront + S3 + Lambda)
Lambda@Edge is also supported (this example uses Viewer Request). Calculated with 10 million accesses and 80% cache hit rate.
Closing
I plan to continue operating this as a personal project, with more services to be added. I'm building it as a tool that I myself find useful.
Cost estimation for serverless architectures is surprisingly complex when you factor in implicit costs and request propagation. I hope Flowdget helps you quickly grasp the overall cost picture while designing your architecture!
For accurate estimates, please use AWS Pricing Calculator . Flowdget is an approximation tool intended for the early stages of architecture planning.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article