AWS

Agents on AWS#

• Marc Brooker breaks down the core components of agentic systems and Bedrock AgentCore — start here.
• Agent SOPs — the structured agent-prompt format used heavily inside Amazon, open-sourced.
• Strands: Strands Labs — AWS’s experimental agent-development GitHub org · building Strands agents: custom tools and MCP integration · Strands & MCP in the open-protocols series · “Still Not Durable Agents” — the sharp critique of Strands and Agent Framework.
• architecting for agentic AI development on AWS · evaluating AI agents: real-world lessons from Amazon · a conversational troubleshooting assistant for K8s workloads.
• AWS MCP servers, put to work as a solutions architect.
• Earlier gen-AI plumbing: serverless prompt chaining with Bedrock and Step Functions · few-shot prompting and fine-tuning in Bedrock.

Well-Architected & the architecture tradition#

• the Well-Architected Framework as a map — “also usable for other cloud providers,” per your save. There’s even a coloring book.
• Let’s Architect! — the series you kept saving: multi-tenancy · serverless solutions · resiliency patterns and trade-offs · DevOps best practices.
• the AWS Fault Isolation Boundaries whitepaper — “don’t let the boring title fool you.”
• The Architect Elevator — connecting IT and the boardroom (Gregor Hohpe, re) and what top-tier software architects do differently.
• the decision guides — “which service do I actually need?” · the Architecture Center as a hands-on gym.
• 20 questions asked in every AWS Solutions Architect interview · the solution-architect mindset.

Serverless-first#

• “the only code you will have to write is business logic” — becoming a serverless-first engineer · the serverless-first mindset as a delivery advantage.
• Yan Cui’s serverless school · Serverless Picks of the Week — the two subscriptions.
• the pragmatic serverless Python developer and refactoring to serverless (videos).
• structuring a real-life serverless application — beyond the hello-world Lambda.
• advanced integration patterns for loosely coupled systems · event-driven architecture on AWS · GoDaddy’s multi-region event-driven platform.
• states-language.net — the Step Functions state-language spec, bookmarked like the standard it is.

EKS & Kubernetes#

• EKS Blueprints — the production-ideas repo, plus the walkthrough.
• scaling an EKS cluster to 1,000 pods — lessons learned · what a production EKS cluster actually looks like.
• The comparison shelf: EKS vs. AKS vs. GKE and McKinsey deploys the same app to all three and measures.
• halving EKS costs in 15 minutes — the FinOps quick win.

Terraform on AWS#

• Clare Liguori: how Amazon does CI/CD — the essential read; branch management, deployments, and all.
• best practices for Terraform state in CI/CD pipelines · Terraform state management, explained.
• Terragrunt for multi-environment management · terragrunt run-all in CI/CD.
• how Slack uses Terraform — across AWS, DigitalOcean, NS1, and GCP.
• scaling IaC pipelines with Terraform, GitHub Actions, and Proton.

Resilience & chaos#

• how Amazon uses chaos engineering to handle 84K requests per second.
• BMW’s chaos-engineering journey · London Stock Exchange Group’s chaos practice · chaos engineering in incident response.
• The counterweight: DHH’s cloud exit — $1M/year saved leaving AWS · a multi-cloud FinOps CLI claiming 60–75% cost cuts.

Learning & certs#

• Skill Builder’s free training and cert prep · the free-certification rounds · CloudQuest — learn the fundamentals as a game.
• Adrian Cantrill’s courses + the paid-community mini-projects repo.
• The Big IAM Challenge — an IAM CTF — learn the scariest service by attacking it.
• the official workshop catalog · AWS re/Start — the free 12-week program.

All links on this topic#