One pipeline. Near-zero at low traffic.
AWSMindset is a personal field journal that publishes itself. EventBridge runs the RSS pipeline every morning, Amazon Bedrock prepares articles, and DynamoDB publishes them with automatic TTL retention. The stack uses serverless, pay-per-use AWS services.
Published
20
articles live
Retention
15 days
automatic TTL
Sources
9
RSS feeds active
Categories
4
content arcs
The pipeline
From RSS to reader, in five stages
Tap any stage to see how it works.
Curated RSS feeds.
1. Sources
Nine engineering and AWS-focused feeds, ordered by editorial priority. AWS announcements and the AWS blog run first; Netflix, Meta, Pragmatic Engineer, React, web.dev, GitHub Blog, and Hacker News (filtered to 100+ points) fill the time budget.
The stack
Every tool, every cost
Twelve building blocks. Nothing exotic, nothing locked in. Everything picked for “stays free at this volume, swappable when it doesn't”.
Compute
- $0 · OSS
App Router, RSC, ISR
- Pay per use
Edge hosting + Next.js compute
- $0 · bundled
Node 20
Runtime for cron + OG image gen
Data
- Pay per use
Posts, subscribers, contact messages
- $0 · public
RSS feeds
9 curated engineering sources
AI
- Usage based
Grounded fact packs + six editorial templates
Comms
- Pay per email
Newsletter delivery
Frontend
- $0 · OSS
CSS-first config, @theme tokens
- $0 · OSS
Command palette (⌘K)
- $0 · built-in
next/og
Per-post OG images at the edge
- $0 · OSS
react-markdown
GFM article rendering
Pattern library
Decisions you can reuse
Practical AWS playbooks with request flows, security boundaries, failure modes, cost drivers, and production checklists.
Separate agent reasoning from side effects
Keep model orchestration separate from the Lambda functions that read data or change AWS resources.
Amazon Bedrock · AWS Lambda · IAM
Decouple consumers with event-driven fan-out
Publish domain events once, then let independent consumers process them at their own pace.
Amazon EventBridge · Amazon SQS · AWS Lambda
Put backpressure between producers and workers
Use a durable queue to absorb traffic spikes and let workers scale without overwhelming downstream systems.
Amazon SQS · AWS Lambda · CloudWatch
Keep the synchronous path intentionally small
Use a thin API layer for validation and authorization, then move longer work behind an asynchronous boundary.
Amazon API Gateway · AWS Lambda · DynamoDB
Keep analytical data open and queryable
Store analytical data in an open table format so multiple query engines can share it without repeated copies.
Amazon S3 · Apache Iceberg · Amazon Athena
Scale services around measurable demand
Run long-lived workloads behind a load balancer and scale tasks from demand and saturation signals.
Amazon ECS · AWS Fargate · Elastic Load Balancing
Cache public content at the edge
Serve immutable assets and cacheable responses close to readers while keeping dynamic work behind explicit origins.
Amazon CloudFront · Amazon S3 · AWS Lambda
Make trust boundaries explicit
Assign narrow identities to each workload and encrypt data with ownership and access paths documented.
IAM · AWS KMS · AWS Secrets Manager
Design observability around user impact
Connect service metrics, traces, and structured logs to a small set of workload-level objectives.
Amazon CloudWatch · AWS X-Ray · AWS CloudTrail
Measure unit cost before optimizing spend
Tie AWS charges to a workload outcome so teams can distinguish healthy growth from architectural waste.
AWS Cost Explorer · AWS Budgets · Cost Anomaly Detection
Total monthly cost
Usage based
CloudFront caches static assets and pages, Lambda runs only when requested, DynamoDB charges by usage, and SES charges by recipient. At personal-blog traffic this should stay near zero, though it is usage-based rather than guaranteed free. Keep AWS Budgets enabled.